Source: Deep Learning on Medium
How To Train Your Own SSD Object Detection Model(Based on Tensorflow)?
Anyone who’s interested in training SSD model can refer to the following link:
- Install LabelImg
- Using LabelImg to annotate photos
- Config SSD
- Download pre-trained model
- Train model
- Use the model to predict
After reading the article, I found that the training process of SSD is similar to YOLOv3. They both use LabelImg to annotate photos but with different file type.
For SSD is VOC (*.xml)；for YOLOv3 is YOLO(*.txt)