Added plot to show false detection

README.md

@@ -18,15 +18,4 @@ sdk: gradio
 ## Usage
 - Upload an image
 - Change settings as requied
-- Hit submit and view results
-
-## Features
-- GradCAM: To understand what models has actully learned. Adjust opacity and model layer for grad-cam.
-- Miss classified images: Plot of images missclassified by model
-- Image input/output: Upload image and check predictions
-- Top classes: Show top n classes with high prediction confidence. 
-
-## Usage
-- Upload an image
-- Change settings as requied
-- Hit submit and view results
+- Hit submit and view result

app.py

@@ -1,6 +1,11 @@
+import os
+import cv2
+import math
 import torch
 import numpy as np
 import gradio as gr
+import matplotlib.pyplot as plt
+from glob import glob
 from PIL import Image
 from models.common import DetectMultiBackend
 from utils.augmentations import letterbox
@@ -14,6 +19,9 @@ data = "data.yaml"
 device = select_device('cpu')
 model = DetectMultiBackend(weights, device=device, dnn=False, data=data, fp16=False)
 
+false_detection_data = glob(os.path.join("false_detection", '*.jpg'))
+false_detection_data = [x.replace('\\', '/') for x in false_detection_data]
+
 def resize_image_pil(image, new_width, new_height):
 
     # Convert to PIL image
@@ -35,7 +43,22 @@ def resize_image_pil(image, new_width, new_height):
 
     return resized
 
-def inference(input_img, conf_thres, iou_thres):
+def display_false_detection_data(false_detection_data, number_of_samples):
+    fig = plt.figure(figsize=(10, 10))
+
+    x_count = 5
+    y_count = 1 if number_of_samples <= 5 else math.floor(number_of_samples / x_count)
+
+    for i in range(number_of_samples):
+        plt.subplot(y_count, x_count, i + 1)
+        img = cv2.imread(false_detection_data[i])
+        plt.imshow(img)
+        plt.xticks([])
+        plt.yticks([])
+
+    return fig
+
+def inference(input_img, conf_thres, iou_thres, is_false_detection_images=True, num_false_detection_images=10):
     im0 = input_img.copy()
     stride, names, pt = model.stride, model.names, model.pt
     imgsz = check_img_size((640, 640), s=stride)  # check image size
@@ -76,8 +99,14 @@ def inference(input_img, conf_thres, iou_thres):
                 c = int(cls)  # integer class
                 label = f'{names[c]} {conf:.2f}'
                 annotator.box_label(xyxy, label, color=colors(c, True))
+    
+    if is_false_detection_images:
+        # Plot the misclassified data
+        misclassified_images = display_false_detection_data(false_detection_data, number_of_samples=num_false_detection_images)
+    else:
+        misclassified_images = None
 
-    return im0
+    return im0, misclassified_images
 
 title = "YOLOv9 model to detect shirt/tshirt"
 description = "A simple Gradio interface to infer on YOLOv9 model and detect tshirt in image"
@@ -90,8 +119,11 @@ examples = [["image_1.jpg", 0.25, 0.45], ["image_2.jpg", 0.25, 0.45],
 demo = gr.Interface(inference, 
                     inputs = [gr.Image(width=320, height=320, label="Input Image"), 
                               gr.Slider(0, 1, 0.25, label="Confidence Threshold"),
-                              gr.Slider(0, 1, 0.45, label="IoU Thresold")],
-                    outputs= [gr.Image(width=640, height=640, label="Output")],
+                              gr.Slider(0, 1, 0.45, label="IoU Thresold"),
+                              gr.Checkbox(label="Show False Detection"),
+                              gr.Slider(5, 35, value=10, step=5, label="Number of False Detection")],
+                    outputs= [gr.Image(width=640, height=640, label="Output"), 
+                              gr.Plot(label="False Detection")],
                     title=title,
                     description=description,
                     examples=examples)