Image Retrieval#
Image retrieval refers to the process of finding images that are relevant to a given search query. Traditionally, this task has relied on metadata such as tags, annotations, or textual descriptions. However, more advanced systems perform retrieval based on the visual content of images themselves, enabling searches based on color, texture, shape, or even high-level semantic features.
The goal of this project is to implement a content-based image retrieval system using PyTorch. Specifically, you will develop one of the following applications.
Choice 1: Image retrieval based on color palettes.
Choice 2: Image retrieval based on face attributes.
Choice 3: Image retrieval based on sketches.
Each of these three applications explores a different aspect of visual similarity. They illustrate how image retrieval can be guided by distinct forms of input, from low-level color information to high-level semantic or structural cues. The individual project descriptions presented in the next pages outline the specific objectives, methodologies, and expected challenges of each approach.
Color Image Retrieval |
Face Image Retrieval |
Sketch Image Retrieval |
---|---|---|