25:null 26:["$","title",null,{"children":"Final Project Process Documentation | ychsu6 | Observable"}] 2b:T494,md` ## Experiment 1 - Getting datasets, Grouping attributes categories, and Pairing with annotations #### Purpose: Having an idea of what dataset I have and how I may play around with it ### Found Dataset - CelebAMask-HQ Dataset images (30,000 faces selected by CelebA dataset) & CelebAMask-HQ Dataset Mask annotations (B& W threshold grayscale) *Dataset-> Female: Male = 2:1 - 40* (1, -1 ) attributes CelebA dataset (300,000 faces) - Position annotation (x, y, z) of faces for CelebAMask-HQ Dataset images

Apply CelebAMask-HQ Dataset Mask annotations as alpha value to CelebAMask-HQ Dataset images to return the separated components images.
Approach: Convert image and reSized components mask to PNG by running script in Terminal with [Imagemagick](https://imagemagick.org/index.php) dependency + trim: convert main.jpg mask.jpg -resize 1024x1024 -alpha off -compose copy-opacity -composite -trim +repage new.png `2c:T546,md` ### Group attributes categories Group 40 attributes(1 or -1) into Components Categories (hair, eyes, nose, lips…).
Gender: Male (1 is male; -1 is female)
-- Basic Components Categories - Hair: Black_Hair, Blond_Hair, Straight_Hair, Wavy_Hair, Brown_Hair, Grey_Hair, Bald, Receding_Hairline, Bangs, Sideburns - Eyebrows: Arched_Eyebrows, Bushy_Eyebrows, - Eyes: Narrow_Eyes, Bags_Under_Eyes, Eyeglasses - Lips: Big_Lips, Wearing_Lipstick, Mouth_Slightly_Open, Smiling - Nose: Pointy_Nose, Big_Nose
-- Additional Components Categories (may exclude, depends on visuals) - Face: Heavy_Makeup, Pale_Skin, Oval_Face, High_Cheekbones, Rosy_Cheeks, Double_Chin (translated as clothes/color?) - Adjective: Attractive, Young, Chubby, Blurry (translated to color/geometrical form?) - Additional: Goatee, Mustache, No_Beard, 5_o_Clock_Shadow (gender-specific so perhaps exclude?) - Accessories: Wearing_Necklace, Wearing_Necktie, Wearing_Earrings, Wearing_Hat (gender-specific so perhaps exclude?) ### Pair Components Categories with annotations - Hair: hair - Eyebrows: l_brow, r_brow - Eyes: l_eye, r_eye - Lips: u_lip, l_lip, mouth - Nose: nose ? Exclude: ear_r(earrings), cloth, hat, ear, neck, skin, l_ear, r_ear `2d:T69a,md` ## Experiement 2 - Cubism template and Facial Components Compositions, Demo on Photoshop #### Purpose: Experimenting the visual outputs Select one Cubism composition template with corresponding landmark points
Early- stage Visual References:

### Possible Approaches for creating Cubism template and Facial Components Compositions #### Experiment then decide how to get the template, whether to: - Extract and get the center point/scale/angle data from existing cubism works to create landmark points to align - Comparing male and female Cubism artists' portrait work -> see from different genders' point of view (why always male artists' POV?) - Manually create arbitrary components composition (adobe PS) and organize my approaches into rules - Create different grids of each attributes, align them in order (according to x, y, z position or size from small to big...), list out when selecting manually, or can even print out/display them with the final work
#### Experiment then decides how to create components visuals, whether to: - Decide which components to duplicate the numbers, which not (2 noses, 2 lips, 2 eyes? 4 eyes?) - Cut out in rectangular form to look more like cubism (function to get the smallest rectangle that contains the form, by cutting the square from the component annotation mask) - Add additional geometrical forms (translated from Face and adjective categories?) to look like cubism `2e:T50e,md` ### Different manual approaches, Demo on Photoshop #### 1st and 2nd Approaches (Number and Shape experiment) -> fail

#### Observations - Main reason for failing -> need a face background!! - "Noses" are what help create "multiple perspectives" -> always try to get 1 frontal + 1 (the most)side; side nose should be placed at edge (according to pose annotation notes) - The order of layering does matter -> hair, face, then other components; both lips then mouth - Perhaps don't cut in rectangular shapes as it's time-consuming to generate new rectangular masks and then again apply all back - Stick to 2 eyes, 2 eyebrows, 2 noses, 2 mouths (4 lips), 2 hair images. Why only eyes are not duplicated? Portraits with more than two eyes tend to make viewers feel passively being observed, instead of being the active observer while looking at two eyes portrait (applies to bother my examples and the photo collage references I included on top), but I wish to remain the "objectified" or "being observed" feelings in my work. `2f:Ta64,md` #### 3rd approach (background faces added + Edge Detection Experiment) 2 huge faces (1 female + 1 male) as background -> Apply precise cut down (not rectangular) facial components on top -> Blur + Apply Edge Detection filters to eliminate the color/lightness differences + Reduce saturation
Demonstrated in PS (original - edgeDetec - edgDetecHue):

#### References: [How to apply Edge Detection on images in p5js](https://idmnyu.github.io/p5.js-image/Edge/index.html) #### Observations: - Background faces have to be one frontal and one side views, but the remaining facial components can be random. Cuz in most cubism works, it's always a combination of two facial outlines that help create such "multi-perspectives" feeling. - Hair should be placed below faces - In general, should highlight the size differences and orientation differences (just aesthetically) - Perhaps don't need additional geometrical forms to highlight "cubism", as it's just the "multi-perspectives" concept I'm referring to, and I wish the focus should be facial components. - It's difficult to generate templates for exact (x, y) positions and sizes of individual components, as they come in different sizes, angles, and shapes. However, if only apply them according to annotated center points, the unexpected offsets in visuals in fact seem interesting. (by 12, 12, 2021)
#### To-do: Create more templates and transfer them into automate-able data(x,y positions of center points, scales(?, pose/perspectives range for later visual displays) ------- `30:Tb3e,md` ## Experiment 3 - Style Transfer with Cubists #### Purpose: Having a quick idea for how the final work may look like by mockups via Photoshop and Style Transfer Testing webpage, Style Transfer Researching According to Wiki Art, among all well-known cubism artists less than 10% of them are female, while in fact the majority of subjects matter in cubism works are female figures. Male artists renowned as Picasso was for being a serial philanderer; however, the stories behind the faces of his lovers in his frames are considerably less well known.
Besides, the way a subject is portrayed by an artist shapes the viewer’s perception of that subject, giving the artist power over the subject, so if women are characterized as a passive object of beauty, whose primary role is to merely look desirable, this thus becomes the strongest idea associated with them, leading to unconscious bias. Connecting to the basic idea of Style Transfer, that: it is possible to separate the Style representation and Content representations in a CNN. Therefore, I wish to separate the Contents (Well-known Celebrities) and Styles (Various Perspectives from both male and female Cubists) in my works, in contrast to those unknown content and uniform bias in the history of Cubism.
(Of course, the order of my approaches was in fact reversed (Machine learning -> Feminism research -> Find dataset -> Female Cubists -> Style transfer), but I found it surprisingly interesting to connect those dots altogether.)
### Male and Female Cubists #### Male Artists and their Cubism works - [Pablo Picasso](https://www.wikiart.org/en/pablo-picasso/all-works#!#filterName:Style_cubism,resultType:masonry) - [Romare Bearden](https://www.wikiart.org/en/romare-bearden/all-works#!#filterName:Style_cubism,resultType:masonry) - [Marc Chagall](https://www.wikiart.org/en/marc-chagall/all-works#!#filterName:Style_cubism,resultType:masonry) #### Female Artists and their Cubism works - [Marevna Marie Vorobieff](https://www.wikiart.org/en/marevna-marie-vorobieff/all-works#!#filterName:Style_cubism,resultType:masonry) - [Natalia Goncharova](https://www.wikiart.org/en/natalia-goncharova/all-works#!#filterName:Style_cubism,resultType:masonry) - [Marie Laurencin](https://www.wikiart.org/en/marie-laurencin/all-works#!#filterName:Style_cubism,resultType:masonry) - [Tamara de Lempicka](https://www.wikiart.org/en/tamara-de-lempicka/all-works#!#filterName:Style_cubism,resultType:masonry) - [Lyubov Popova](https://www.wikiart.org/en/lyubov-popova/all-works#!#filterName:Style_cubism,resultType:masonry) - [Olga Rozanova](https://www.wikiart.org/en/olga-rozanova/all-works#!#filterName:Style_cubism,resultType:masonry) `31:T692,md` ### More Research on the algorithm and Why I decided to use Style Transfer After getting the first blended version (all generated style looks quite similar), I was wondering whether Style Transfer was just "generalizing", giving an overview, or merely "fetching" color palette and randomly assigning onto images, because strokes, shapes and compositions, (all that make Cubism works different from one and another) are gone. Therefore, I spend a few days digging a little bit more into the algorithm, trying to question whether such style transfer models are sufficient to work as real artists' jobs to generate a "representation" for the subject. After going through Dumoulin, et. al.’s and Ghiasi, et. al.'s papers, I realized how these losses are extracted (the distribution of brush strokes is taken out at some of the early layers while the identification of content is extracted at a much deeper layer), it does match with how I wish the style can be transferred in my work. Plus, I figured that just by simply replacing both the Style Model and the Transformer Model from small ones to big ones, the visual results look significantly better!

#### A little bit more about the algorithm: Since the style-specific parameters allow us to “combine” different styles by taking the weighted average between style representations of multiple styles, it allows some (especially in Ghiasi, et. al.'s work) interpolation between the style representation of arbitrary styles, by calculating the style representation of both the content image and the style image and taking their weighted average. `32:Tbd8,md` Here I use both bigger Style Model + bigger Transformer Model, it works much better than the smaller models:
Original Inceptionv3 style model + Original transformer model #### Intput Content:

#### + Pablo Picasso's style:

#### + Romare Bearden's style:

#### + Marc Chagall's style:

#### + Natalia Goncharova's style:

#### + Lyubov Popova's style:

#### + Marevna Marie Vorobieff's style:

### Style Transfer Demostration [TensorFlow + P5.js testing](https://mlq4l.csb.app/)
#### References for the Code: [Style Trnasfer](https://github.com/reiinakano/arbitrary-image-stylization-tfjs/blob/master/main.js) [Tensorflow API](https://js.tensorflow.org/api/1.7.1)
[Tensorflow.js on Geeks for geeks](https://www.geeksforgeeks.org/tensorflow-js-tf-tensor-function/?ref=gcse)
[JaveScript- Promise](https://javascript.info/promise-basics)

### Observations - The size of the input style image does influence the visual outcomes (the bigger the style image size, the less stylized the output images) - Which style model and transformer model I choose do highly influence the outcomes (by 18, 12, 2021) ------- `33:T4e3,md` Irrelevant but just Personal reflections:
1. Making critical works is haaaaaard!!!! it isn't just about composing a story from what I'm just randomly interested in but must have certain degrees of knowledge on all these topics (at this point I'm still not sure I understand everything perfectly, but I tried my best...). or else all the connections just seems arbitrary and superficial. Especially with works wishing to make certain reflections about the tools themselves, then every single step or selection of tools must have enough reasoning supported behind, or else again the argument just seems very... weak :( 2. For the first time de-cluttering thoughts is killing me... Even all these ideas didn't seem like irrelevant dots to me in earlier stages, but after realizing that too many crazy ideas were going on and such focus point was missing, then trying to reconnect everything and making sure nothing is too arbitrary was driving me crazy... (don't understand why such divergent-then-de-cluttering process always work in the past tho...)
Tho I'm sure even if I start from a smaller and simpler idea next time, my crazy minds might again spread through all the topics and greedily want to include so many ideas into one... :( `34:[{"id":"57ce6afc1c11403f8ea3a1522553226e99e30989b0b368f182053fc737de0aca17ecb1614022f8805debc6e342d8487da0b9617944f560c708d939c65b0d6148","href":"https://static.observableusercontent.com/files/57ce6afc1c11403f8ea3a1522553226e99e30989b0b368f182053fc737de0aca17ecb1614022f8805debc6e342d8487da0b9617944f560c708d939c65b0d6148","name":"piccasso-1.png","type":"image/png","lastModified":1639574149672.778,"size":129175},{"id":"e52ddef90ea474e68b2e176efa728a26077e52357fcf0db7822d7f239173b94cd30d28c98092e6f7fec0c23e473a632686fe373c330494d698d06d9a2419ad2a","href":"https://static.observableusercontent.com/files/e52ddef90ea474e68b2e176efa728a26077e52357fcf0db7822d7f239173b94cd30d28c98092e6f7fec0c23e473a632686fe373c330494d698d06d9a2419ad2a","name":"00000_nose@1.png","type":"image/png","lastModified":1639126939027.089,"size":1361},{"id":"dd188a16c02ab10c97150895ddfb9ff2ea977c661ade46e852b44b0a5d2d1f7e3a990a1cff7ed3282d0278e99556994aa32681d2ebb4f1eedec1ecdb9b4be228","href":"https://static.observableusercontent.com/files/dd188a16c02ab10c97150895ddfb9ff2ea977c661ade46e852b44b0a5d2d1f7e3a990a1cff7ed3282d0278e99556994aa32681d2ebb4f1eedec1ecdb9b4be228","name":"00000_nose.png","type":"image/png","lastModified":1639126919557.033,"size":40595},{"id":"816fa4c932a75ee605b5e52627ad73766d50fe70360a92badc16860e482a6e4e5fe5c61f731c4f6492e5c382e723f8896742be1929935ea94af7791c0e5bf617","href":"https://static.observableusercontent.com/files/816fa4c932a75ee605b5e52627ad73766d50fe70360a92badc16860e482a6e4e5fe5c61f731c4f6492e5c382e723f8896742be1929935ea94af7791c0e5bf617","name":"0.jpg","type":"image/jpeg","lastModified":1639126838182.126,"size":160346},{"id":"67591d4d70510355e7f34e6f2511e42c1118ac8d2a22b8fe90ca4c255e65f7de55bc25ab0feb1cb2218d178f7f50e4a8265ae5b708515e07b686e5d4382e6a76","href":"https://static.observableusercontent.com/files/67591d4d70510355e7f34e6f2511e42c1118ac8d2a22b8fe90ca4c255e65f7de55bc25ab0feb1cb2218d178f7f50e4a8265ae5b708515e07b686e5d4382e6a76","name":"Screenshot 2021-12-10 at 5.58.18 PM.png","type":"image/png","lastModified":1639130312381.076,"size":418867},{"id":"0da4049d46c28ed8beec05e2945138a2c92a243894ff0776dbc2a05c91629dbbd5a8c664218598d9556b1a5ce3c7d73d4c7940fc8d1685761b2d35c0c3eda523","href":"https://static.observableusercontent.com/files/0da4049d46c28ed8beec05e2945138a2c92a243894ff0776dbc2a05c91629dbbd5a8c664218598d9556b1a5ce3c7d73d4c7940fc8d1685761b2d35c0c3eda523","name":"fhfjhgg.jpeg.crdownload","type":"image/jpeg","lastModified":1639133890727.471,"size":50665},{"id":"c5e46e00a8166b70c98fc7146f9903b78b0ea0d2f28c70f00971f305f2a8905173cc5d6acb983ed015dd6dc2c3d9b4e791fc6cb68c68b449f6afb17be4c36141","href":"https://static.observableusercontent.com/files/c5e46e00a8166b70c98fc7146f9903b78b0ea0d2f28c70f00971f305f2a8905173cc5d6acb983ed015dd6dc2c3d9b4e791fc6cb68c68b449f6afb17be4c36141","name":"asdgfg.jpeg","type":"image/jpeg","lastModified":1639133890435.252,"size":5858},{"id":"e83a05de3e62fc20ac65a3edcf156d4e2cc7b7e86cdd50da6984c80c3c20050653ddf9369994c61f0f63c38d0edeac5bf3246d9abd698c81ede2433ffc6dedfd","href":"https://static.observableusercontent.com/files/e83a05de3e62fc20ac65a3edcf156d4e2cc7b7e86cdd50da6984c80c3c20050653ddf9369994c61f0f63c38d0edeac5bf3246d9abd698c81ede2433ffc6dedfd","name":"fdf.jpeg","type":"image/jpeg","lastModified":1639133890424.604,"size":5105},{"id":"0b2297d6d5e43e9d379382ed62501835d3ae229549c94645e1cea74eaa57ed10ee370488046bba9b4852cb71a1a30be98248f69e47c6f185ea240f060ee3f2c3","href":"https://static.observableusercontent.com/files/0b2297d6d5e43e9d379382ed62501835d3ae229549c94645e1cea74eaa57ed10ee370488046bba9b4852cb71a1a30be98248f69e47c6f185ea240f060ee3f2c3","name":"messageImage_1639327312409.jpg","type":"image/jpeg","lastModified":1639336287970.129,"size":108924},{"id":"d63eca3c4eaf57a705ca8443a0c996cc359fca92e28ea9d65817f21a13ef299ff59bc212089e3672f095582824fd5dbd93944e3a507118216b633ac9d9f28170","href":"https://static.observableusercontent.com/files/d63eca3c4eaf57a705ca8443a0c996cc359fca92e28ea9d65817f21a13ef299ff59bc212089e3672f095582824fd5dbd93944e3a507118216b633ac9d9f28170","name":"messageImage_1639327223973.jpg","type":"image/jpeg","lastModified":1639336287829.421,"size":83934},{"id":"9e55ee876579b70e8f73723d11a92bf4ae1867141cb3552661cbaf0282597f619dd91abcc48fe8003ae05c3c596c230026f5d4bb33bdde48852fd62893b1a8cf","href":"https://static.observableusercontent.com/files/9e55ee876579b70e8f73723d11a92bf4ae1867141cb3552661cbaf0282597f619dd91abcc48fe8003ae05c3c596c230026f5d4bb33bdde48852fd62893b1a8cf","name":"tryOut-1-edgeDetec.png","type":"image/png","lastModified":1639347205539.16,"size":761030},{"id":"59b03bc21138255f1914ed19c7a3e5c8ce697d47dba9dc64945d2c9770e8651cb36d803d0f7e9b243d836531b8d64ae23708b4ad2821ca10388f5974030f0d82","href":"https://static.observableusercontent.com/files/59b03bc21138255f1914ed19c7a3e5c8ce697d47dba9dc64945d2c9770e8651cb36d803d0f7e9b243d836531b8d64ae23708b4ad2821ca10388f5974030f0d82","name":"tryOut-1.png","type":"image/png","lastModified":1639347205473.886,"size":710395},{"id":"84326856cf42794d20f663aaf92ddcc301b09b4ebdc5cada31dc646c33ce1810bcad2d3ef8cdf56a1328edbc8c6036a324cac68d16bc8be776302b7ff8ab1bf2","href":"https://static.observableusercontent.com/files/84326856cf42794d20f663aaf92ddcc301b09b4ebdc5cada31dc646c33ce1810bcad2d3ef8cdf56a1328edbc8c6036a324cac68d16bc8be776302b7ff8ab1bf2","name":"tryOut-1 -edgeDetecHue.png","type":"image/png","lastModified":1639347205155.743,"size":311991},{"id":"8449b8ca1571de4bc98b9dd1a09e50bc465a4c373230aa84cf8d484edbe431d137871cbc83513395fa70147a13af284febff23514dad8a3574d3b6a4fa61a925","href":"https://static.observableusercontent.com/files/8449b8ca1571de4bc98b9dd1a09e50bc465a4c373230aa84cf8d484edbe431d137871cbc83513395fa70147a13af284febff23514dad8a3574d3b6a4fa61a925","name":"tryOut-2(1).png","type":"image/png","lastModified":1639353535701.288,"size":675207},{"id":"7d0b00e38dcd551857b33cead719761177328efae758f89ee200c0d3a808639921e4ce8d02b7f4fd9457ed938eb52c187be3e9904f293622d5dcde27bb4d58c6","href":"https://static.observableusercontent.com/files/7d0b00e38dcd551857b33cead719761177328efae758f89ee200c0d3a808639921e4ce8d02b7f4fd9457ed938eb52c187be3e9904f293622d5dcde27bb4d58c6","name":"tryOut-2(1) -edgeDetec.png","type":"image/png","lastModified":1639353535620.542,"size":724555},{"id":"2c6707cce309a1af9e392552fafc661bd946146066146d434f0f4f3c411525e899e8cafdfcc38c03ba3aae39631d010bc1ee0a11492958723c1ce0cd252339bc","href":"https://static.observableusercontent.com/files/2c6707cce309a1af9e392552fafc661bd946146066146d434f0f4f3c411525e899e8cafdfcc38c03ba3aae39631d010bc1ee0a11492958723c1ce0cd252339bc","name":"tryOut-2(1) -edgeDetecHue.png","type":"image/png","lastModified":1639353535287.707,"size":300231},{"id":"a274f3f5b8c663b214309b792d10b8c970e061e750ad9e4d14cb8ebd63a39d6c5b27d509642b80af7a5cf00bc7a8f6422023d7d63ff86cf8f014fe29f5b44a5b","href":"https://static.observableusercontent.com/files/a274f3f5b8c663b214309b792d10b8c970e061e750ad9e4d14cb8ebd63a39d6c5b27d509642b80af7a5cf00bc7a8f6422023d7d63ff86cf8f014fe29f5b44a5b","name":"tryOut-3 -edgeDetec.png","type":"image/png","lastModified":1639353296695.087,"size":710622},{"id":"fde3993163719a25386dd91f1d19c73a7ef604c41a1678ee89c63e313aa8679bec4209c3d35ee88f16599f526bcbeea8030661b68621628365a9134443cb7374","href":"https://static.observableusercontent.com/files/fde3993163719a25386dd91f1d19c73a7ef604c41a1678ee89c63e313aa8679bec4209c3d35ee88f16599f526bcbeea8030661b68621628365a9134443cb7374","name":"tryOut-3.png","type":"image/png","lastModified":1639353296623.394,"size":598485},{"id":"40fc8833c30eb5c94f83b21a209382705d4f1cfb87733a9562afa032fa6340ad2f8a53bb2cc4e63b883b73957014999f90c6b99eb4866806f843fb0b4bd79204","href":"https://static.observableusercontent.com/files/40fc8833c30eb5c94f83b21a209382705d4f1cfb87733a9562afa032fa6340ad2f8a53bb2cc4e63b883b73957014999f90c6b99eb4866806f843fb0b4bd79204","name":"tryOut-3 -edgeDetecHue.png","type":"image/png","lastModified":1639353296291.201,"size":290209},{"id":"f5801b1f3d8745595a1d38af0ce9f606fd46c708cff79ccec687e67aa5455fa62aad0a9757165483813844ea6fc339940b9b297c18014ea54ad7eb615197e95a","href":"https://static.observableusercontent.com/files/f5801b1f3d8745595a1d38af0ce9f606fd46c708cff79ccec687e67aa5455fa62aad0a9757165483813844ea6fc339940b9b297c18014ea54ad7eb615197e95a","name":"Marc Chagall-1.jpg","type":"image/jpeg","lastModified":1639578447372.27,"size":15836},{"id":"11233f45060da3ba66e3f5b29f6beb3743ea10776a5ced6290edc7c1d2ab4bd0498d0d559c394c1ab45a4e413918cd08a5bd3afe25e0cd378883183a7ee03c69","href":"https://static.observableusercontent.com/files/11233f45060da3ba66e3f5b29f6beb3743ea10776a5ced6290edc7c1d2ab4bd0498d0d559c394c1ab45a4e413918cd08a5bd3afe25e0cd378883183a7ee03c69","name":"tryOut-4 -edgeDetec@1.png","type":"image/png","lastModified":1639583031952.577,"size":744951},{"id":"2a74f07cfffe11bc8a90e7f03cafd561ae403f4203451e60df91629b2d3b47d82b35ad3be70d53b15d1cb33fae4b0ef8d248b3f7cd48882ec82569df4b892f14","href":"https://static.observableusercontent.com/files/2a74f07cfffe11bc8a90e7f03cafd561ae403f4203451e60df91629b2d3b47d82b35ad3be70d53b15d1cb33fae4b0ef8d248b3f7cd48882ec82569df4b892f14","name":"tryOut-4@1.png","type":"image/png","lastModified":1639583031845.176,"size":636477},{"id":"46339d3698b54bc065278ad5b414a0b5b52857d082592399a117cd500e6b15496bf6e2931984657df5fc9a09a2058ffb81920f16dd1bd4ac5b81db597fde3f0c","href":"https://static.observableusercontent.com/files/46339d3698b54bc065278ad5b414a0b5b52857d082592399a117cd500e6b15496bf6e2931984657df5fc9a09a2058ffb81920f16dd1bd4ac5b81db597fde3f0c","name":"tryOut-4 -edgeDetecHue@1.png","type":"image/png","lastModified":1639583031420.804,"size":306309},{"id":"859b2282fa0b226fd3ee04004f14531d87f7d581fd8329707c541f46b7a18444068aadde877f7a0576a9f7abbfe56b4e934d634b0f46b0b662a19d7e653782e0","href":"https://static.observableusercontent.com/files/859b2282fa0b226fd3ee04004f14531d87f7d581fd8329707c541f46b7a18444068aadde877f7a0576a9f7abbfe56b4e934d634b0f46b0b662a19d7e653782e0","name":"Romare Bearden-6.jpg","type":"image/jpeg","lastModified":1639827705867.416,"size":19147},{"id":"353a126cb7084b1e604e355f902b7d7860cdba3226af08e1c03779a18a8ab64c30890a796aaff17fc47222b13370d368f480ef65d81b271b33194a65cd337e17","href":"https://static.observableusercontent.com/files/353a126cb7084b1e604e355f902b7d7860cdba3226af08e1c03779a18a8ab64c30890a796aaff17fc47222b13370d368f480ef65d81b271b33194a65cd337e17","name":"+Marevna Marie Vorobieff-2Style.png","type":"image/png","lastModified":1639827997057.052,"size":118853},{"id":"54f4910b892b4a875adc49a48fbd7465623b76e4bf42a2adfe30c4b4efdaa5748d5758eb7003b5ba3f79bf58cdeab36a1d52c88c485469fc874c472ef0f61071","href":"https://static.observableusercontent.com/files/54f4910b892b4a875adc49a48fbd7465623b76e4bf42a2adfe30c4b4efdaa5748d5758eb7003b5ba3f79bf58cdeab36a1d52c88c485469fc874c472ef0f61071","name":"Marevna Marie Vorobieff-2.jpg","type":"image/jpeg","lastModified":1639827989644.533,"size":14476},{"id":"3590666758e94e616f2ffe647061d2d33c561867a4730629f93d888545b5b5f41e6b93eb2764fb826baccafbbb8da8cb315e3c10763503b96209fde5b62cbd53","href":"https://static.observableusercontent.com/files/3590666758e94e616f2ffe647061d2d33c561867a4730629f93d888545b5b5f41e6b93eb2764fb826baccafbbb8da8cb315e3c10763503b96209fde5b62cbd53","name":"+Romare Bearden-6Style.png","type":"image/png","lastModified":1639827678864.019,"size":132864},{"id":"dcc5858c40128645bb830bb1c0ad921e9509d1c68c5c07b7a6f38314542265dff496e5f5d2d4f14e3f498b8c200ad960c45aa92ac38aec9e63ee8cd83a8562a2","href":"https://static.observableusercontent.com/files/dcc5858c40128645bb830bb1c0ad921e9509d1c68c5c07b7a6f38314542265dff496e5f5d2d4f14e3f498b8c200ad960c45aa92ac38aec9e63ee8cd83a8562a2","name":"Lyubov Popova-4.jpg","type":"image/jpeg","lastModified":1639827461205.148,"size":21384},{"id":"e9af972eebea6517458d2773f25e68942f1d25a448e56e90d717b954c00f8f05308a053c40a8a0df68a35f56ffc8ff7219a85fb5986c6db0067fbd076ac4d790","href":"https://static.observableusercontent.com/files/e9af972eebea6517458d2773f25e68942f1d25a448e56e90d717b954c00f8f05308a053c40a8a0df68a35f56ffc8ff7219a85fb5986c6db0067fbd076ac4d790","name":"+Lyubov Popova-4Style.png","type":"image/png","lastModified":1639827443032.913,"size":129980},{"id":"0f53f9b46d24037a404ba1c5dff47d0dd8115ce368dfeb697874063569b33a444faaa50e950b1453b7998a9d943d7f58db1bca04dc46a1a416b0d4915daaa437","href":"https://static.observableusercontent.com/files/0f53f9b46d24037a404ba1c5dff47d0dd8115ce368dfeb697874063569b33a444faaa50e950b1453b7998a9d943d7f58db1bca04dc46a1a416b0d4915daaa437","name":"Natalia Goncharova-6.jpg","type":"image/jpeg","lastModified":1639827398035.278,"size":12409},{"id":"cd87bc89acd5cd5bfa9bb0173b4bb613141600411e49ec2797f94e3861aab2eddc9259000b3b45b7da4a52313062e6703b14d22432771f4ca26115471cf05f83","href":"https://static.observableusercontent.com/files/cd87bc89acd5cd5bfa9bb0173b4bb613141600411e49ec2797f94e3861aab2eddc9259000b3b45b7da4a52313062e6703b14d22432771f4ca26115471cf05f83","name":"+Natalia Goncharova-6Style.png","type":"image/png","lastModified":1639827373046.455,"size":118891},{"id":"e84511aaca93910ea49370bf00c367d82e7af461408921f2871810ca8f19c2c96bc5135242ec9d65ebf5fa7e16e63a41daabed698b5da47169d498b5c514a9e6","href":"https://static.observableusercontent.com/files/e84511aaca93910ea49370bf00c367d82e7af461408921f2871810ca8f19c2c96bc5135242ec9d65ebf5fa7e16e63a41daabed698b5da47169d498b5c514a9e6","name":"+Marc Chagall-1Style.png","type":"image/png","lastModified":1639827281482.362,"size":121242},{"id":"dac02cb093811e4488e82c84742a9114eccf9177ca4d6a5c2ff80c0f38f724bda9f374b9fd0df3536493d1a3746cf4e03880c742cf2a20309ddadf877a9ee83e","href":"https://static.observableusercontent.com/files/dac02cb093811e4488e82c84742a9114eccf9177ca4d6a5c2ff80c0f38f724bda9f374b9fd0df3536493d1a3746cf4e03880c742cf2a20309ddadf877a9ee83e","name":"+piccasso-1Style.png","type":"image/png","lastModified":1639827271336.45,"size":128280},{"id":"eeff9391bf1c6100834b857cf55e0aebf726190bcbd95845e9b934461eefbfb6357970d94a5819a45766c3c9f54bbb50cb733e33a8baf5fc7598873586220637","href":"https://static.observableusercontent.com/files/eeff9391bf1c6100834b857cf55e0aebf726190bcbd95845e9b934461eefbfb6357970d94a5819a45766c3c9f54bbb50cb733e33a8baf5fc7598873586220637","name":"++Romare Bearden-6Style-2.png","type":"image/png","lastModified":1639873223760.306,"size":134275},{"id":"2fb3fcfe7f5c40961450b38cb9942cbfa2e3061808987fbc485a9d9b45210d2114d7547abeb3a82a164d3e70a6c8b1bae27fc6f52c466affb63ad5a483f29b8d","href":"https://static.observableusercontent.com/files/2fb3fcfe7f5c40961450b38cb9942cbfa2e3061808987fbc485a9d9b45210d2114d7547abeb3a82a164d3e70a6c8b1bae27fc6f52c466affb63ad5a483f29b8d","name":"+Lyubov Popova-4Style-2.png","type":"image/png","lastModified":1639873223679.535,"size":130544},{"id":"55df562ef61fe23435e1d5e43e2e735f529cc38c9c594e511833bfff694e7347d2cff032b211e6070b784a5aa8a33881b06f10836c98877e50cb1363a4c9b81e","href":"https://static.observableusercontent.com/files/55df562ef61fe23435e1d5e43e2e735f529cc38c9c594e511833bfff694e7347d2cff032b211e6070b784a5aa8a33881b06f10836c98877e50cb1363a4c9b81e","name":"+Natalia Goncharova-6Style-2.png","type":"image/png","lastModified":1639873223638.797,"size":118662},{"id":"140dee1239b136673b7c76e8094f6b1c53a431945fd84f8dd2d7a8e1e74d4fb3503a0057ebe1d4532d5b54284f34178aaccf6347cfcad8319321c2ec67ab4c8e","href":"https://static.observableusercontent.com/files/140dee1239b136673b7c76e8094f6b1c53a431945fd84f8dd2d7a8e1e74d4fb3503a0057ebe1d4532d5b54284f34178aaccf6347cfcad8319321c2ec67ab4c8e","name":"+picasso-1Style-2.png","type":"image/png","lastModified":1639873223711.147,"size":129438},{"id":"5c565afb743811fa4f6649c69b8affa2e614768634b0dbdd384c97c455aca727c4319c4ad6140e64a722af9032fa76700c494f94d095a9b76c9c88a512eb92a5","href":"https://static.observableusercontent.com/files/5c565afb743811fa4f6649c69b8affa2e614768634b0dbdd384c97c455aca727c4319c4ad6140e64a722af9032fa76700c494f94d095a9b76c9c88a512eb92a5","name":"+Marc Chagall-1Style-2.png","type":"image/png","lastModified":1639873223660.224,"size":122071},{"id":"32f134d084f5a18052138ef9b196fe1bcc67cf641e398411ad1f276c669cb0012420e6a4718a568f5e7060fc7880246c963677c09049612f7984fcfc0efaa1de","href":"https://static.observableusercontent.com/files/32f134d084f5a18052138ef9b196fe1bcc67cf641e398411ad1f276c669cb0012420e6a4718a568f5e7060fc7880246c963677c09049612f7984fcfc0efaa1de","name":"+Marevna Marie Vorobieff-2Style-2.png","type":"image/png","lastModified":1639873223647.088,"size":119821},{"id":"7c9aae93730dc7bbfe98c8d9f3e56bee0611bb2a9f8cd40da1ca84a366fc0fea0e63606563ac9e7b285c908aba851324bfed5b6f77807b3438ab71c5e1b1be3b","href":"https://static.observableusercontent.com/files/7c9aae93730dc7bbfe98c8d9f3e56bee0611bb2a9f8cd40da1ca84a366fc0fea0e63606563ac9e7b285c908aba851324bfed5b6f77807b3438ab71c5e1b1be3b","name":"faces.png","type":"image/png","lastModified":1639877080484.069,"size":1597655},{"id":"6e58d9738541df164a2abe59deeabc0b927674b320ea2423d670de1c56ede88201b6f4111a0dcc26ff353e410ab932ee38a73c62a42782e01bf8ea2c183a5bb8","href":"https://static.observableusercontent.com/files/6e58d9738541df164a2abe59deeabc0b927674b320ea2423d670de1c56ede88201b6f4111a0dcc26ff353e410ab932ee38a73c62a42782e01bf8ea2c183a5bb8","name":"VGG.png","type":"image/png","lastModified":1640030173626.745,"size":117413}] 29:{"id":"a5f453318c44b45a","slug":"final-project/2","title":"Final Project Process Documentation","thumbnail":"32e50a65c19b7473b0a481e7562a3fc952468e49b70c965ac21a5419745372c2","access_level":"public","create_time":"$D2021-12-10T08:12:36.921Z","update_time":"$D2021-12-20T20:10:34.432Z","view_time":null,"first_public_time":"$D2021-12-10T10:01:30.105Z","has_unpublished_edits":null,"fork":false,"stars":0,"starred":false,"owner":{"type":"team","id":"c3d91d7db2867b4d","login":"betahsu","name":"ychsu6","avatar_url":"https://avatars-next.observableusercontent.com/197b7792c36f626b9fe391a472486b8b6dc3cf3503a2bb52e365e7916c324a9e","tier":"starter_2024"},"version":747,"publish_version":747,"role":null,"authored":false,"authors":[{"type":"individual","id":"8205ffe96d046f8c","login":"betahsu","name":"ychsu6","avatar_url":"https://avatars-next.observableusercontent.com/197b7792c36f626b9fe391a472486b8b6dc3cf3503a2bb52e365e7916c324a9e","tier":null}],"members":[{"type":"individual","id":"8205ffe96d046f8c","login":"betahsu","name":"ychsu6","avatar_url":"https://avatars-next.observableusercontent.com/197b7792c36f626b9fe391a472486b8b6dc3cf3503a2bb52e365e7916c324a9e","tier":null,"role":"editor"}],"comments":[],"custom_thumbnail":false,"trashed":false,"lapsed":null,"subscribed":null,"forks":0,"collection":null,"type":"notebook","body":{"title":"Final Project Process Documentation","cells":[{"id":0,"value":"# Final Project Process Documentation","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"","database":"$undefined"},{"id":6,"value":"$2b","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":73,"value":"$2c","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":68,"value":"md`\n### Pairing \"Scores\"\n

\nAlgorithm each time randomly select a male and one attribute with its “score” (hairCat: 1, 1, 1, 1, -1, -1, -1, 1, 1, -1), then go through all females to pick one attribute (hair) with the same “score” as pairs \n\n(by 10, 12, 2021)\n\n---------------\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":112,"value":"$2d","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":284,"value":"$2e","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":287,"value":"$2f","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":361,"value":"$30","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":736,"value":"md`\n### Sample Visuals for Content + Cubists' [Style Transfer](https://reiinakano.com/arbitrary-image-stylization-tfjs/) \n#### Different Models:\n[High Quality] Original Inceptionv3 style model
\n[Fast] Distilled MobileNet style model
\n[High Quality] Original transformer model
\n[Fast] Separable_conv2d transformer
\n\nOriginally I use both smaller Style Model + smaller Transformer Model, the result looks rather similar:
\nDistilled MobileNet style model + Separable_conv2d transformer
\n

\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":740,"value":"$31","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":513,"value":"$32","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":688,"value":"md`\n#### Fundamental \"tuning\"? for Concepts and Methodology\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":604,"value":"md`\nSince the idea has again changed quite a bit, I'll leave this page as a documentation for how the ideas have evolved.
\nHere's another (hopefully) more organized documentation version:\n## [Official Final Project Page](https://observablehq.com/@betahsu/final-project/3)\n\n-------\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":677,"value":"$33","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":584,"value":"md`\n\n`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"}],"files":"$W34","stdlib":"1","titling":"auto"}} 2a:"$29:body"