Object Detection Workshop
===
## What Is YOLO and What Can It do?
:::info
Official Resources
- Website: https://docs.ultralytics.com
- Github: https://github.com/ultralytics/ultralytics
:::
YOLO (You Only Look Once) - a popular object detection model
| Sports | Safety |
| -------- | -------- |
| <img alt="Football Player Detection" src="https://github.com/RizwanMunawar/ultralytics/assets/62513924/7d320e1f-fc57-4d7f-a691-78ee579c3442"> | <img alt="People Fall Detection" src="https://github.com/RizwanMunawar/ultralytics/assets/62513924/86437c4a-3227-4eee-90ef-9efb697bdb43"> |
| Football Player Detection| People Fall Detection |
## Detection with Official Pre-trained Model
- Reference:
- Ultralytics: https://docs.ultralytics.com
- Github: https://github.com/ultralytics/ultralytics
- Arguements: https://docs.ultralytics.com/modes/predict/#inference-arguments
- Workshop Notebook: https://colab.research.google.com/drive/1BVvbLD3z3Z88EVO991RT3zfc3wm4hcwc
### Install Packages
Download the package from your terminal
```python=
pip install ultralytics
```
If you use colab notebook, add this line into your cell
```shell=
!pip install ultralytics
```
### Load Pre-trained Model
```python=
from ultralytics import YOLO
model = YOLO('yolov8n.pt') # yolov8n.pt 即是預訓練好的模型
```
:::info
you can choose different pre-trained model
- `yolov8n.pt`
- `yolov8s.pt`
- `yolov8m.pt`
- `yolov8l.pt`
- `yolov8x.pt`
:::
### Inference on an image
you should see the `/runs/detect/predict/bus.jpg` in your working directory
```python=
source = 'https://ultralytics.com/images/bus.jpg'
results = model(source, save=True, conf=0.5)
```
| Original | Inference|
| -------- | -------- |
| <img src="https://hackmd.io/_uploads/H1JTj1mLp.jpg)" width=405, height=540/>| <img src="https://hackmd.io/_uploads/r1naiJX86.jpg" width=405, height=540/> |
### Inference on a video
```python=
source = 'https://github.com/LittleFish-Coder/object-detection-workshop/raw/master/newjeans.mp4'
results = model(source, save=True, conf=0.5)
```
<img src="https://github.com/LittleFish-Coder/object-detection-workshop/raw/master/src/newjeans.gif" alt="newjeans">
### Select Certain Classes
#### COCO Dataset and Classes
The [COCO](https://cocodataset.org/#home) (Common Objects in Context) dataset is a large-scale object detection, segmentation, and captioning dataset. It is designed to encourage research on a wide variety of object categories and is commonly used for benchmarking computer vision models. It is an essential dataset for researchers and developers working on object detection, segmentation, and pose estimation tasks.
Pre-trained模型是在COCO dataset上預先做訓練的，而coco dataset擁有80種不同的objects(classes)
:::spoiler 所有object classes
0: person
1: bicycle
2: car
3: motorcycle
4: airplane
5: bus
6: train
7: truck
8: boat
9: traffic light
10: fire hydrant
11: stop sign
12: parking meter
13: bench
14: bird
15: cat
16: dog
17: horse
18: sheep
19: cow
20: elephant
21: bear
22: zebra
23: giraffe
24: backpack
25: umbrella
26: handbag
27: tie
28: suitcase
29: frisbee
30: skis
31: snowboard
32: sports ball
33: kite
34: baseball bat
35: baseball glove
36: skateboard
37: surfboard
38: tennis racket
39: bottle
40: wine glass
41: cup
42: fork
43: knife
44: spoon
45: bowl
46: banana
47: apple
48: sandwich
49: orange
50: broccoli
51: carrot
52: hot dog
53: pizza
54: donut
55: cake
56: chair
57: couch
58: potted plant
59: bed
60: dining table
61: toilet
62: tv
63: laptop
64: mouse
65: remote
66: keyboard
67: cell phone
68: microwave
69: oven
70: toaster
71: sink
72: refrigerator
73: book
74: clock
75: vase
76: scissors
77: teddy bear
78: hair drier
79: toothbrush
:::
#### Select Certain Classes (person)
```python=
classes = [0] # 0: person, 1: bicycle, 2: car, 3: motorcycle, 5: airplane ...
source = 'https://ultralytics.com/images/bus.jpg'
results = model(source, save=True, conf=0.5, classes=classes)
```
#### Count How Many Objects In The Picture
```python=
classes = [0] # 0: person, 1: bicycle, 2: car, 3: motorcycle, 4: airplane, 5: bus ...
source = 'https://ultralytics.com/images/bus.jpg'
results = model.predict(source, save=True, conf=0.5, classes=classes)
for r in results:
print(r.boxes.cls) # r.boxes.cls 為預測結果的類別
# tensor([0., 0., 0.) # 3個0代表預測結果為3個person
source = 'https://ultralytics.com/images/bus.jpg'
results = model.predict(source, save=True, conf=0.5, classes=None)
for r in results:
print(r.boxes.cls) # r.boxes.cls 為預測結果的類別
# tensor([5., 0., 0., 0.) # 5代表預測結果為airplane, 3個0代表預測結果為3個person
# count number of people
source = 'https://ultralytics.com/images/bus.jpg'
results = model.predict(source, save=True, conf=0.5, classes=None)
people = 0
for r in results:
for c in r.boxes.cls:
print(model.names[int(c)])
if int(c) == 0:
people += 1
print(f'There are {people} people') # 3
```
### Inference On Your Webcam (會有機率當機)
- Install OpenCV
```shell=
pip install opencv-python
```
- CV2.VideoCapture()
```pytohn=
import cv2
video_path = 0 # 0 代表使用電腦內建的攝影機
cap = cv2.VideoCapture(video_path)
# Loop through the video frames
while cap.isOpened():
# Read a frame from the video
success, frame = cap.read()
if success:
# Run YOLOv8 inference on the frame
results = model(frame)
# Visualize the results on the frame
annotated_frame = results[0].plot()
# Display the annotated frame
cv2.imshow("YOLOv8 Inference", annotated_frame)
# Break the loop if 'q' is pressed
if cv2.waitKey(1) & 0xFF == ord("q"):
break
else:
# Break the loop if the end of the video is reached
break
# Release the video capture object and close the display window
cap.release()
cv2.destroyAllWindows()
```
## Customize Model
- Official Tutorial: https://youtu.be/a3SBRtILjPI
- Label Platform: https://roboflow.com (Use Chrome Browser!)
- Dataset Universe: https://universe.roboflow.com/
### Prepare Your Own Dataset
Label your own dataset
### Export yaml file
The file contains
data.yml is the meta data of your dataset
### Transfer Learning
```python=
# train the model on custom dataset
yml_path = 'data.yaml'
results = model.train(data=yml_path, epochs=10) # the result will be saved in runs/detect/train/weights
```
Sign in with Wallet
Connect another wallet