YOLO v4 and TF 2.0

Topics covered

1. Installing Anaconda Python

2. Creating a virtual python environments

Topics covered

1. Covid-19 chest X-ray images dataset preparation

2. Resizing dataset for training

Topics covered

1. Image labeling using labelimg

2. Identifying the best techniques for image labeling

Topics covered

1. Converting YOLO format to PASCAL VOC format for TF2.0 SSD(Single Shot Detector) training.

2. Understanding the difference between these formats.

Topics covered

1. YOLOv4 training overview.

2. YOLOv4 weights to tflite conversion.

3. What is quantization and how it speed things!

Topic covered

1. Deploying the converted model as an android application

Topics covered

1. SSD mobilenet training overview

2. SSD mobilenet weights to tflite conversion

Topic covered

1. TF2.0 android deployment using the trained model

Topics covered

Basics

1. Precision  and Recall

2. IoU(Intersection Over Union)

3. Mean Average Precision/Average Precision(mAP/AP)

4. Batch Normalization

5. Residual blocks

6. Activation function

7. Max pooling

8. Feature Pyramid Networks(FPN)

9. Path Aggregation Network (PAN)

10. SPP (spatial pyramid pooling layer)

11. Channel Attention Module(CAM) and Spatial Attention Module (SAM)

YOLOv4 - Technical details

1. Backbone

2. Cross-Stage-Partial-connections (CSP)

3. YOLO with SPP

4. PAN in YOLOv4

5. Spatial Attention Module (SAM) in YOLOv4

6. Bag of freebies (Bof) and Bag of specials (BoS)

SSD - Technical details

1. Architecture overview and working

2. Loss functions

YOLO vs SSD

1. Speed and accuracy benchmarking

Summary

Reference