Andrew Zisserman, Ragav Sachdeva
We have not lifted any functions out of this paper's repositories yet, so there is nothing we have run. The repositories linked to it are listed below.
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
The scene shows a young character[0] sitting cross-legged on a bed in a relaxed, casual bedroom setting. The bed has a rumpled look, with a soft blanket and pillow, giving a sense of comfort and informality. The character [0], dressed in a checkered shirt over a T-shirt, is holding up a small object, possibly a fan or similar item, with a focused expression. Nearby, a few personal items lie scattered, including a book or notebook, adding to the lived-in feel of the room. The setting suggests a private moment of contemplation or introspection in a cozy, familiar environment. The scene captures a dynamic moment as a character[0] swiftly extends their[0] arm, suggesting a quick or forceful action, possibly throwing or releasing something with intent. The movement is accentuated by motion lines trailing from the arm, conveying speed and energy. The character's[0] expression appears focused and determined, with their[0] body slightly turned to enhance the sense of movement. The checkered shirt flares with the motion, adding to the visual impact of the gesture. The overall scene emphasizes action and concentration, highlighting the character's[0] physical engagement in the moment. The scene shows two characters in a domestic setting, likely a kitchen area. One character[0], with short hair and wearing a patterned shirt, stands behind a counter and appears to be speaking to the other person[1], who has longer hair and is seen from behind. The counter is decorated with a bowl of what seems to be fruit or a similar item, along with a small plant in a vase and a cup, adding to the homely atmosphere. In the background, there is a tiled wall with a large kitchen hood, cabinets, and other typical kitchen features, giving a sense of warmth and familiarity to the space. In this scene, a tall, imposing apartment building stretches upward under a darkened sky, with its balconies casting sharp shadows that emphasize the structure's height. Inside one of the apartments, two characters interact in a cozy kitchen setting. Daichi's mother, standing behind a counter filled with small homely items like a fruit bowl and a vase, is inquiring about her son, Daichi. Yuki, a young person with long hair, responds that Daichi is likely in his room. 1 By literary we mean that the output is text-based. calising panels, characters, texts, and speech-bubble tails, performing OCR, grounding characters etc. (2) We release human-annotated captions for over 3300 Japanese comic panels, along with character grounding annotations, and benchmark large vision-language models in their ability to understand comic images. (3) Finally, we demonstrate how integrating large vision-language models with Magiv3, can generate seamless literary narratives that allows visually impaired audiences to engage with the depth and richness of comic storytelling. Our code, trained model and dataset annotations can be found at: https://github.com/ ragavsachdeva/magi.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2503.23344")
get_code_for_paper("2503.23344")
have("2503.23344")
The run record, dated, one paper per request, free:
curl https://syntology.ai/api/ran/2503.23344.json
A badge for a README (the split and the date, never a ratio):
[](https://syntology.ai/paper/2503.23344)
Connect an agent — have() is free.