# Computer-Vision-Relative-Attributes

**Repository Path**: atomai/Computer-Vision-Relative-Attributes

## Basic Information

- **Project Name**: Computer-Vision-Relative-Attributes
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2020-12-01
- **Last Updated**: 2020-12-20

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

CS676A Project - Group 1
========
**Vikas Jain - 13788**
**Shubham Jain - 13683**    

### Langauges:
Python and Matlab

### Papers followed:
1. Parikh, Devi, and Kristen Grauman. "Relative attributes." Computer Vision (ICCV), 2011 IEEE International Conference on. IEEE, 2011.
2. Burges, Chris, et al. "Learning to rank using gradient descent." Proceedings of the 22nd international conference on Machine learning. ACM, 2005.

3. Joachims, Thorsten. "Optimizing search engines using clickthrough data." Proceedings of the eighth ACM SIGKDD international conference on Knowledge discovery and data mining. ACM, 2002.

4. Simonyan, Karen, and Andrew Zisserman. "Very deep convolutional networks for large-scale image recognition." arXiv preprint arXiv:1409.1556 (2014).

### Dataset:
PubFig Dataset  
"Attribute and Simile Classifiers for Face Verification,"
Neeraj Kumar, Alexander C. Berg, Peter N. Belhumeur, and Shree K. Nayar,
International Conference on Computer Vision (ICCV), 2009.

### Code
Our project consisted of three parts - **feature extraction** and **two ranking models** (RankSVM and RankNet).  
Each part has separate folder in _code/_ directory with their separate README file.  
See README file of each part for the dependencies and commands to execute the code.  

1. **Feature extraction** -  in code/cnn folder. It is implemented in keras. The details are in the readme in that folder.

2. **Ranking using RankSVM** - in code/RankSVM folder. It is implemented in matlab. The details are in the readme in that folder.

3. **Ranking using RankNet** - in code/RankNet folder. It is implemented in python. The details are in the readme in that folder.