wenxiaowei / food-recognition-project

Repo for Deep Learning project

food-recognition-project

Repo for Deep Learning project

The Food Recognition challenge is an image segmentation problem aiming to recognize individual food items in each image.

The dataset contains:

• A training set of 24120 food images, along with their corresponding 39328 annotations in MS-COCO format, and comprising proper segmentation, classification, and volume/weights estimates.
• A Validation and Test sets of 1269 images.

Evaluation Criteria

Given a known ground truth mask A, and a predicted mask B, first compute Intersection Over Union (IoU). The prediction is tradionally considered a True detection, when there is at least half an overlap, i.e. IoU >0.5. Then you may define precision and recall. The final scoring parameters are computed by averaging over all the precision and recall.
The final scoring parameters are computed by averaging over all the precision and recall values for all known annotations in the ground truth.

See also this discussion.

Project development

We decide to implement this project by using "One hundred layers Tiramisu", for more details, you may want to read the report

Setup

In order to make be able to run the project jupyter notebook, you should have the data set in the folder containing the script organized in this way:
├───annotations
├───images
│ ├───test
│ ├───train
│ └───val
├─── weights
├─── script file
├─── other files