centroids-reid

CTL-Model : On the Unreasonable Effectiveness of Centroids in Image Retrieval

Input (query)

(Image from http://188.138.127.15:81/Datasets/Market-1501-v15.09.15.zip)

Shape : (batch, 3, height, width)

Output (Top 10 images from gallery images that are similar to query)

Usage

Automatically downloads the onnx and prototxt files on the first run. It is necessary to be connected to the Internet while downloading.

For the sample image,

$ python3 centroids-reid.py

If you want to specify the input image, put the image path after the --input option.
You can use --savepath option to change the name of the output file to save.

$ python3 centroids-reid.py --input IMAGE_PATH --savepath SAVE_IMAGE_PATH

If you want to specify the directory of gallery image, put the directory path after the --gallery_dir option.

$ python3 centroids-reid.py --gallery_dir gallery

Now, files in this gallery directory are very restricted.
Many more files can be found in the bounding_box_test directory of Market-1501-v15.09.15.zip.

By adding the model name after the --model option, you can specify the model.
The model name is selected from 'market1501_resnet50', 'dukemtmcreid_resnet50'.

$ python3 centroids-reid.py --model market1501_resnet50

Reference

• On the Unreasonable Effectiveness of Centroids in Image Retrieval

Framework

Pytorch

Model Format

ONNX opset=11

Netron

market1501_resnet50_256_128_epoch_120.onnx.prototxt
dukemtmcreid_resnet50_256_128_epoch_120.onnx.prototxt