https://github.com/katerynazakharova/olive-separation-with-tensorflow-object-detection-api

Object detection on a custom dataset
https://github.com/katerynazakharova/olive-separation-with-tensorflow-object-detection-api

Last synced: 4 months ago
JSON representation

Object detection on a custom dataset

• Host: GitHub
• URL: https://github.com/katerynazakharova/olive-separation-with-tensorflow-object-detection-api
• Owner: KaterynaZakharova
• Created: 2021-01-20T14:01:55.000Z (over 4 years ago)
• Default Branch: main
• Last Pushed: 2021-01-20T15:09:44.000Z (over 4 years ago)
• Last Synced: 2025-01-31T22:57:35.365Z (5 months ago)
• Language: Jupyter Notebook
• Size: 14.3 MB
• Stars: 2
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Olive-Separation-With-TensorFlow-Object-Detection-API

The task is to identify and to count black and green olives in the video.

The steps:

### 1. Make 75 photos.

(25 black olives, 25 green olives, 25 both)

**Examples of images:**

![image_example_1](image_example_1.jpg)

![image_example_2](image_example_2.jpg)

![image_example_3](image_example_3.jpg)

### 2. Transform images into a size of 300x300.

Using *transform_image_resolution.py*

### 3. Create annotations in LableImg.

Creating CSV format from XML using *xml_to_csv.py*

### 4. Generate TFRecords.

Using *generate_tfrecord.py*

### 5. Create a label map file and a configuration file.

The used configuration file is https://github.com/tensorflow/models/blob/master/research/object_detection/configs/tf2/ssd_efficientdet_d0_512x512_coco17_tpu-8.config

### 6. Train model.

The used model is http://download.tensorflow.org/models/object_detection/tf2/20200711/efficientdet_d0_coco17_tpu-32.tar.gz

### 7. Export the inference graph.

### 8. Test the model.

**Input:**

![frame0](frame0.jpg)

![frame70](frame70.jpg)

**Output:**

![output_frame0](output_frame0.png)

![output_frame70](output_frame70.png)

`8 black olives.`

`1 green olives.`