PawMatchAI / app.py
DawnC's picture
Update app.py
d7ed8db verified
raw
history blame
21.7 kB
import os
import numpy as np
import torch
import torch.nn as nn
import gradio as gr
from torchvision.models import efficientnet_v2_m, EfficientNet_V2_M_Weights
from torchvision.ops import nms, box_iou
import torch.nn.functional as F
from torchvision import transforms
from PIL import Image, ImageDraw, ImageFont, ImageFilter
from data_manager import get_dog_description
from urllib.parse import quote
from ultralytics import YOLO
import asyncio
import traceback
model_yolo = YOLO('yolov8l.pt')
dog_breeds = ["Afghan_Hound", "African_Hunting_Dog", "Airedale", "American_Staffordshire_Terrier",
"Appenzeller", "Australian_Terrier", "Bedlington_Terrier", "Bernese_Mountain_Dog",
"Blenheim_Spaniel", "Border_Collie", "Border_Terrier", "Boston_Bull", "Bouvier_Des_Flandres",
"Brabancon_Griffon", "Brittany_Spaniel", "Cardigan", "Chesapeake_Bay_Retriever",
"Chihuahua", "Dandie_Dinmont", "Doberman", "English_Foxhound", "English_Setter",
"English_Springer", "EntleBucher", "Eskimo_Dog", "French_Bulldog", "German_Shepherd",
"German_Short-Haired_Pointer", "Gordon_Setter", "Great_Dane", "Great_Pyrenees",
"Greater_Swiss_Mountain_Dog", "Ibizan_Hound", "Irish_Setter", "Irish_Terrier",
"Irish_Water_Spaniel", "Irish_Wolfhound", "Italian_Greyhound", "Japanese_Spaniel",
"Kerry_Blue_Terrier", "Labrador_Retriever", "Lakeland_Terrier", "Leonberg", "Lhasa",
"Maltese_Dog", "Mexican_Hairless", "Newfoundland", "Norfolk_Terrier", "Norwegian_Elkhound",
"Norwich_Terrier", "Old_English_Sheepdog", "Pekinese", "Pembroke", "Pomeranian",
"Rhodesian_Ridgeback", "Rottweiler", "Saint_Bernard", "Saluki", "Samoyed",
"Scotch_Terrier", "Scottish_Deerhound", "Sealyham_Terrier", "Shetland_Sheepdog",
"Shih-Tzu", "Siberian_Husky", "Staffordshire_Bullterrier", "Sussex_Spaniel",
"Tibetan_Mastiff", "Tibetan_Terrier", "Walker_Hound", "Weimaraner",
"Welsh_Springer_Spaniel", "West_Highland_White_Terrier", "Yorkshire_Terrier",
"Affenpinscher", "Basenji", "Basset", "Beagle", "Black-and-Tan_Coonhound", "Bloodhound",
"Bluetick", "Borzoi", "Boxer", "Briard", "Bull_Mastiff", "Cairn", "Chow", "Clumber",
"Cocker_Spaniel", "Collie", "Curly-Coated_Retriever", "Dhole", "Dingo",
"Flat-Coated_Retriever", "Giant_Schnauzer", "Golden_Retriever", "Groenendael", "Keeshond",
"Kelpie", "Komondor", "Kuvasz", "Malamute", "Malinois", "Miniature_Pinscher",
"Miniature_Poodle", "Miniature_Schnauzer", "Otterhound", "Papillon", "Pug", "Redbone",
"Schipperke", "Silky_Terrier", "Soft-Coated_Wheaten_Terrier", "Standard_Poodle",
"Standard_Schnauzer", "Toy_Poodle", "Toy_Terrier", "Vizsla", "Whippet",
"Wire-Haired_Fox_Terrier"]
class MultiHeadAttention(nn.Module):
def __init__(self, in_dim, num_heads=8):
super().__init__()
self.num_heads = num_heads
self.head_dim = max(1, in_dim // num_heads)
self.scaled_dim = self.head_dim * num_heads
self.fc_in = nn.Linear(in_dim, self.scaled_dim)
self.query = nn.Linear(self.scaled_dim, self.scaled_dim)
self.key = nn.Linear(self.scaled_dim, self.scaled_dim)
self.value = nn.Linear(self.scaled_dim, self.scaled_dim)
self.fc_out = nn.Linear(self.scaled_dim, in_dim)
def forward(self, x):
N = x.shape[0]
x = self.fc_in(x)
q = self.query(x).view(N, self.num_heads, self.head_dim)
k = self.key(x).view(N, self.num_heads, self.head_dim)
v = self.value(x).view(N, self.num_heads, self.head_dim)
energy = torch.einsum("nqd,nkd->nqk", [q, k])
attention = F.softmax(energy / (self.head_dim ** 0.5), dim=2)
out = torch.einsum("nqk,nvd->nqd", [attention, v])
out = out.reshape(N, self.scaled_dim)
out = self.fc_out(out)
return out
class BaseModel(nn.Module):
def __init__(self, num_classes, device='cuda' if torch.cuda.is_available() else 'cpu'):
super().__init__()
self.device = device
self.backbone = efficientnet_v2_m(weights=EfficientNet_V2_M_Weights.IMAGENET1K_V1)
self.feature_dim = self.backbone.classifier[1].in_features
self.backbone.classifier = nn.Identity()
self.num_heads = max(1, min(8, self.feature_dim // 64))
self.attention = MultiHeadAttention(self.feature_dim, num_heads=self.num_heads)
self.classifier = nn.Sequential(
nn.LayerNorm(self.feature_dim),
nn.Dropout(0.3),
nn.Linear(self.feature_dim, num_classes)
)
self.to(device)
def forward(self, x):
x = x.to(self.device)
features = self.backbone(x)
attended_features = self.attention(features)
logits = self.classifier(attended_features)
return logits, attended_features
num_classes = 120
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
model = BaseModel(num_classes=num_classes, device=device)
checkpoint = torch.load('best_model_81_dog.pth', map_location=torch.device('cpu'))
model.load_state_dict(checkpoint['model_state_dict'])
# evaluation mode
model.eval()
# Image preprocessing function
def preprocess_image(image):
# If the image is numpy.ndarray turn into PIL.Image
if isinstance(image, np.ndarray):
image = Image.fromarray(image)
# Use torchvision.transforms to process images
transform = transforms.Compose([
transforms.Resize((224, 224)),
transforms.ToTensor(),
transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
])
return transform(image).unsqueeze(0)
def get_akc_breeds_link(breed):
base_url = "https://www.akc.org/dog-breeds/"
breed_url = breed.lower().replace('_', '-')
return f"{base_url}{breed_url}/"
async def predict_single_dog(image):
image_tensor = preprocess_image(image)
with torch.no_grad():
output = model(image_tensor)
logits = output[0] if isinstance(output, tuple) else output
probabilities = F.softmax(logits, dim=1)
topk_probs, topk_indices = torch.topk(probabilities, k=3)
top1_prob = topk_probs[0][0].item()
topk_breeds = [dog_breeds[idx.item()] for idx in topk_indices[0]]
# Calculate relative probabilities for display
raw_probs = [prob.item() for prob in topk_probs[0]]
sum_probs = sum(raw_probs)
relative_probs = [f"{(prob/sum_probs * 100):.2f}%" for prob in raw_probs]
return top1_prob, topk_breeds, relative_probs
async def detect_multiple_dogs(image, conf_threshold=0.3, iou_threshold=0.45):
results = model_yolo(image, conf=conf_threshold, iou=iou_threshold)[0]
dogs = []
boxes = []
for box in results.boxes:
if box.cls == 16: # COCO dataset class for dog is 16
xyxy = box.xyxy[0].tolist()
confidence = box.conf.item()
boxes.append((xyxy, confidence))
if not boxes:
dogs.append((image, 1.0, [0, 0, image.width, image.height]))
else:
nms_boxes = non_max_suppression(boxes, iou_threshold)
for box, confidence in nms_boxes:
x1, y1, x2, y2 = box
w, h = x2 - x1, y2 - y1
x1 = max(0, x1 - w * 0.05)
y1 = max(0, y1 - h * 0.05)
x2 = min(image.width, x2 + w * 0.05)
y2 = min(image.height, y2 + h * 0.05)
cropped_image = image.crop((x1, y1, x2, y2))
dogs.append((cropped_image, confidence, [x1, y1, x2, y2]))
return dogs
def non_max_suppression(boxes, iou_threshold):
keep = []
boxes = sorted(boxes, key=lambda x: x[1], reverse=True)
while boxes:
current = boxes.pop(0)
keep.append(current)
boxes = [box for box in boxes if calculate_iou(current[0], box[0]) < iou_threshold]
return keep
def calculate_iou(box1, box2):
x1 = max(box1[0], box2[0])
y1 = max(box1[1], box2[1])
x2 = min(box1[2], box2[2])
y2 = min(box1[3], box2[3])
intersection = max(0, x2 - x1) * max(0, y2 - y1)
area1 = (box1[2] - box1[0]) * (box1[3] - box1[1])
area2 = (box2[2] - box2[0]) * (box2[3] - box2[1])
iou = intersection / float(area1 + area2 - intersection)
return iou
async def process_single_dog(image):
top1_prob, topk_breeds, relative_probs = await predict_single_dog(image)
# Case 1: Low confidence - unclear image or breed not in dataset
if top1_prob < 0.15:
initial_state = {
"explanation": "The image is unclear or the breed is not in the dataset. Please upload a clearer image of a dog.",
"image": None,
"is_multi_dog": False
}
return initial_state["explanation"], None, initial_state
breed = topk_breeds[0]
# Case 2: High confidence - single breed result
if top1_prob >= 0.45:
description = get_dog_description(breed)
formatted_description = format_description(description, breed)
initial_state = {
"explanation": formatted_description,
"image": image,
"is_multi_dog": False
}
return formatted_description, image, initial_state
# Case 3: Medium confidence - show top 3 breeds with relative probabilities
else:
breeds_info = ""
for i, (breed, prob) in enumerate(zip(topk_breeds, relative_probs)):
description = get_dog_description(breed)
formatted_description = format_description(description, breed)
breeds_info += f"\n\nBreed {i+1}: **{breed}** (Confidence: {prob})\n{formatted_description}"
initial_state = {
"explanation": breeds_info,
"image": image,
"is_multi_dog": False
}
return breeds_info, image, initial_state
# async def predict(image):
# if image is None:
# return "Please upload an image to start.", None, None
# try:
# if isinstance(image, np.ndarray):
# image = Image.fromarray(image)
# dogs = await detect_multiple_dogs(image)
# color_list = ['#FF0000', '#00FF00', '#0000FF', '#FFFF00', '#00FFFF', '#FF00FF', '#800080', '#FFA500']
# annotated_image = image.copy()
# draw = ImageDraw.Draw(annotated_image)
# font = ImageFont.load_default()
# dogs_info = ""
# for i, (cropped_image, detection_confidence, box) in enumerate(dogs):
# color = color_list[i % len(color_list)]
# draw.rectangle(box, outline=color, width=3)
# draw.text((box[0] + 5, box[1] + 5), f"Dog {i+1}", fill=color, font=font)
# top1_prob, topk_breeds, relative_probs = await predict_single_dog(cropped_image)
# combined_confidence = detection_confidence * top1_prob
# dogs_info += f'<div class="dog-info" style="border-left: 5px solid {color}; margin-bottom: 20px; padding: 15px;">'
# dogs_info += f'<h2>Dog {i+1}</h2>'
# if combined_confidence < 0.2:
# dogs_info += "<p>The image is unclear or the breed is not in the dataset. Please upload a clearer image.</p>"
# elif top1_prob >= 0.45:
# breed = topk_breeds[0]
# description = get_dog_description(breed)
# dogs_info += format_description_html(description, breed)
# else:
# dogs_info += "<h3>Top 3 possible breeds:</h3>"
# for breed, prob in zip(topk_breeds, relative_probs):
# description = get_dog_description(breed)
# dogs_info += f"<div class='breed-section'>"
# dogs_info += f"<h4>{breed} (Confidence: {prob})</h4>"
# dogs_info += format_description_html(description, breed)
# dogs_info += "</div>"
# dogs_info += '</div>'
# html_output = f"""
# <style>
# .dog-info {{
# border: 1px solid #ddd;
# margin-bottom: 20px;
# padding: 15px;
# border-radius: 5px;
# box-shadow: 0 2px 5px rgba(0,0,0,0.1);
# }}
# .dog-info h2 {{
# background-color: #f0f0f0;
# padding: 10px;
# margin: -15px -15px 15px -15px;
# border-radius: 5px 5px 0 0;
# }}
# .breed-section {{
# margin-bottom: 20px;
# padding: 10px;
# background-color: #f8f8f8;
# border-radius: 5px;
# }}
# </style>
# {dogs_info}
# """
# initial_state = {
# "dogs_info": dogs_info,
# "image": annotated_image,
# "is_multi_dog": len(dogs) > 1,
# "html_output": html_output
# }
# return html_output, annotated_image, initial_state
# except Exception as e:
# error_msg = f"An error occurred: {str(e)}\n\nTraceback:\n{traceback.format_exc()}"
# print(error_msg)
# return error_msg, None, None
async def predict(image):
if image is None:
return "Please upload an image to start.", None, None
try:
if isinstance(image, np.ndarray):
image = Image.fromarray(image)
dogs = await detect_multiple_dogs(image)
# 更新為更容易區分的顏色組合
color_list = ['#FF3B30', '#34C759', '#007AFF', '#FF9500', '#5856D6', '#FF2D55', '#5AC8FA', '#FFCC00']
annotated_image = image.copy()
draw = ImageDraw.Draw(annotated_image)
# 改用更大的字體,提升可讀性
try:
font = ImageFont.truetype("arial.ttf", 24) # 優先使用 Arial
except:
font = ImageFont.load_default()
dogs_info = ""
for i, (cropped_image, detection_confidence, box) in enumerate(dogs):
color = color_list[i % len(color_list)]
# 增強框線可見度
draw.rectangle(box, outline=color, width=4)
# 優化標籤背景
label = f"Dog {i+1}"
label_bbox = draw.textbbox((0, 0), label, font=font)
label_width = label_bbox[2] - label_bbox[0]
label_height = label_bbox[3] - label_bbox[1]
# 添加標籤背景
label_x = box[0] + 5
label_y = box[1] + 5
draw.rectangle(
[label_x - 2, label_y - 2, label_x + label_width + 4, label_y + label_height + 4],
fill='white',
outline=color,
width=2
)
draw.text((label_x, label_y), label, fill=color, font=font)
top1_prob, topk_breeds, relative_probs = await predict_single_dog(cropped_image)
combined_confidence = detection_confidence * top1_prob
# 優化資訊卡片樣式
dogs_info += f'''
<div class="dog-info-card" style="border-left: 6px solid {color};">
<div class="dog-info-header" style="background-color: {color}20;">
<span class="dog-label" style="color: {color};">Dog {i+1}</span>
</div>
'''
if combined_confidence < 0.15:
dogs_info += '''
<div class="warning-message">
<p>The image is unclear or the breed is not in the dataset. Please upload a clearer image.</p>
</div>
'''
elif top1_prob >= 0.45:
breed = topk_breeds[0]
description = get_dog_description(breed)
dogs_info += f'''
<div class="breed-info">
<div class="confidence-score" style="color: {color};">
{breed} (Confidence: {relative_probs[0]})
</div>
{format_description_html(description, breed)}
</div>
'''
else:
dogs_info += "<h3>Top 3 possible breeds:</h3>"
for breed, prob in zip(topk_breeds, relative_probs):
description = get_dog_description(breed)
dogs_info += f'''
<div class="breed-section">
<div class="confidence-score" style="color: {color};">
{breed} (Confidence: {prob})
</div>
{format_description_html(description, breed)}
</div>
'''
dogs_info += '</div>'
# 更新 CSS 樣式
html_output = f"""
<style>
.dog-info-card {{
border: 1px solid #ddd;
margin-bottom: 24px;
padding: 0;
border-radius: 8px;
box-shadow: 0 2px 8px rgba(0,0,0,0.1);
overflow: hidden;
transition: all 0.3s ease;
}}
.dog-info-card:hover {{
box-shadow: 0 4px 12px rgba(0,0,0,0.15);
}}
.dog-info-header {{
padding: 16px 20px;
margin: 0;
font-size: 20px;
font-weight: bold;
}}
.dog-label {{
font-size: 18px;
font-weight: bold;
}}
.breed-info {{
padding: 20px;
}}
.breed-section {{
margin: 20px;
padding: 16px;
background-color: #f8f8f8;
border-radius: 6px;
}}
.confidence-score {{
font-size: 18px;
font-weight: bold;
margin-bottom: 12px;
}}
.warning-message {{
padding: 16px 20px;
color: #ff3b30;
font-weight: bold;
}}
</style>
{dogs_info}
"""
initial_state = {
"dogs_info": dogs_info,
"image": annotated_image,
"is_multi_dog": len(dogs) > 1,
"html_output": html_output
}
return html_output, annotated_image, initial_state
except Exception as e:
error_msg = f"An error occurred: {str(e)}\n\nTraceback:\n{traceback.format_exc()}"
print(error_msg)
return error_msg, None, None
def show_details_html(choice, previous_output, initial_state):
if not choice:
return previous_output, gr.update(visible=True), initial_state
try:
breed = choice.split("More about ")[-1]
description = get_dog_description(breed)
formatted_description = format_description_html(description, breed)
html_output = f"""
<div class="dog-info">
<h2>{breed}</h2>
{formatted_description}
</div>
"""
initial_state["current_description"] = html_output
initial_state["original_buttons"] = initial_state.get("buttons", [])
return html_output, gr.update(visible=True), initial_state
except Exception as e:
error_msg = f"An error occurred while showing details: {e}"
print(error_msg)
return f"<p style='color: red;'>{error_msg}</p>", gr.update(visible=True), initial_state
def format_description_html(description, breed):
html = "<ul style='list-style-type: none; padding-left: 0;'>"
if isinstance(description, dict):
for key, value in description.items():
html += f"<li style='margin-bottom: 10px;'><strong>{key}:</strong> {value}</li>"
elif isinstance(description, str):
html += f"<li>{description}</li>"
else:
html += f"<li>No description available for {breed}</li>"
html += "</ul>"
akc_link = get_akc_breeds_link(breed)
html += f'<p><a href="{akc_link}" target="_blank">Learn more about {breed} on the AKC website</a></p>'
return html
with gr.Blocks() as iface:
gr.HTML("<h1 style='text-align: center;'>🐶 Dog Breed Classifier 🔍</h1>")
gr.HTML("<p style='text-align: center;'>Upload a picture of a dog, and the model will predict its breed and provide detailed information!</p>")
gr.HTML("<p style='text-align: center; color: #666; font-size: 0.9em;'>Note: This is an AI model and predictions may not always be 100% accurate. The model provides its best estimates based on training data.</p>")
with gr.Row():
input_image = gr.Image(label="Upload a dog image", type="pil")
output_image = gr.Image(label="Annotated Image")
output = gr.HTML(label="Prediction Results")
initial_state = gr.State()
input_image.change(
predict,
inputs=input_image,
outputs=[output, output_image, initial_state]
)
gr.Examples(
examples=['Border_Collie.jpg', 'Golden_Retriever.jpeg', 'Saint_Bernard.jpeg', 'French_Bulldog.jpeg', 'Samoyed.jpg'],
inputs=input_image
)
gr.HTML('For more details on this project and other work, feel free to visit my GitHub <a href="https://github.com/Eric-Chung-0511/Learning-Record/tree/main/Data%20Science%20Projects/Dog_Breed_Classifier">Dog Breed Classifier</a>')
if __name__ == "__main__":
iface.launch()