Classification and Segmentation 4 Vietnamese Foods

Summary

• In this project, I use Pytorch to perform classification and segmentation of 4 popular dishes in Vietnam.Also, to make image segmentation easy, I used additional library Segmentation Pytorch

Dataset 🥚

• The data is taken from the dataset 30VNFoods, this dataset includes 30 dishes and I took 4 dishes: Bánh mì, Phở, Bánh tráng nướng và Cơm tấm

  

Models

• I use a variety of models ranging from MLP to Simple CNN, miniVGG. Pre-trained models such as VGG16, ResNet18

• For the Segmentation problem, I use the Unet structure with the encoders being pre-trained models to get the best results.

• I use Wandb to track and compare experiments: Classification, Segmentation

How to run this project ❓

git clone https://github.com/Harly-1506/4VNfoods-Deep-learning.git

cd 4VNfoods_Deep_learning
#run classification
run classifi_main.py
#run segmentation
run seg_main.py

Note: When you run seg_main.py, it takes 8 to 10 minutes to prepare the data

Classification Results

Methods	Accuracy	Loss	Val_Accuracy	Val_Loss	Test_accuracy
Resnet18_pretrained	99.926	6.78E-05	96.907	0.1106	95.886
Resnet18	99.486	0.0003	80.154	0.7141	78.663
VGG16_pretrained	99.266	0.0005	94.587	0.4035	95.758
VGG16	95.229	0.0030	78.350	0.6939	77.763
miniVGG	99.926	0.0001	82.989	0.6325	87.917
SimpleCNN	99.559	0.0008	86.597	0.3855	86.632
MLP_4hidden512node	53.651	0.0678	45.103	2.8904	47.043
MLP_3hidden1024node	44.403	0.1080	34.278	4.8297	38.946
MLP_3hidden512node	55.486	0.0707	40.721	5.5563	44.987
MLP_4hidden	47.706	0.0583	37.886	2.3706	38.303
MLP_3hidden	49.761	0.0512	36.082	3.0187	41.902
MLP_2hidden	48.844	0.0438	40.979	1.6916	41.516

Segmentation Results

Methods	iou/valid	iou banhmi	iou banhtrang	iou comtam	iou pho	iou_clutter
Unet_ResNet34	0.8625	0.8273	0.8529	0.7083	0.7099	0.9084
Unet-ResNet18	0.8828	0.8655	0.8897	0.7893	0.7571	0.9214
Unet-VGG16	0.8716	0.8627	0.8713	0.7395	0.7463	0.9146

Plot Val Accuracy

• Classification:
• Segmentation:

Demo:

• Demo program you can follow in this repository: Demo

---

Author: Harly

If you have any problems, please leave a message in Issues

Give me a star ⭐ if you find it useful, thanks