Deepface - 仅以超过阈值的置信度显示面部的情绪

我有这段代码可以检测面部表情，但它会在没有面部表情的地方找到面部表情。所以我想知道是否可以放心地做到这一点。

我尝试在谷歌上搜索如何做到这一点，但没有找到任何有用的东西。

import cv2
from deepface import DeepFace

# Load face cascade classifier
face_cascade = cv2.CascadeClassifier(cv2.data.haarcascades + 'haarcascade_frontalface_default.xml')

# Start capturing video
cap = cv2.VideoCapture('ap.mp4')

while True:
    # Capture frame-by-frame
    ret, frame = cap.read()

    # Convert frame to grayscale
    gray_frame = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)

    # Convert grayscale frame to RGB format
    rgb_frame = cv2.cvtColor(gray_frame, cv2.COLOR_GRAY2RGB)

    # Detect faces in the frame
    faces = face_cascade.detectMultiScale(gray_frame, scaleFactor=1.1, minNeighbors=5, minSize=(30, 30))
    for (x, y, w, h) in faces:
        # Extract the face ROI (Region of Interest)
        face_roi = rgb_frame[y:y + h, x:x + w]

        # Perform emotion analysis on the face ROI
        result = DeepFace.analyze(face_roi, actions=['emotion'], enforce_detection=False)

        # Determine the dominant emotion
        print(result[0]['emotion'])
        emotion = result[0]['dominant_emotion']

        # Draw rectangle around face and label with predicted emotion
        cv2.rectangle(frame, (x, y), (x + w, y + h), (0, 0, 255), 2)
        cv2.putText(frame, emotion, (x, y - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.9, (0, 0, 255), 2)

    # Display the resulting frame
    cv2.imshow('Real-time Emotion Detection', frame)

    # Press 'q' to exit
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break

# Release the capture and close all windows
cap.release()
cv2.destroyAllWindows()

可以通过过滤掉置信度低于特定阈值的情绪来增加对 DeepFace 结果的置信度。以下是更新后的代码，其中包含一个添加的阈值：

import cv2
from deepface import DeepFace

# Load face cascade classifier
face_cascade = cv2.CascadeClassifier(cv2.data.haarcascades + 'haarcascade_frontalface_default.xml')

# Start capturing video
cap = cv2.VideoCapture('ap.mp4')

# 设置置信度阈值
confidence_threshold = 0.6  # 例如 60% 的置信度

while True:
    # Capture frame-by-frame
    ret, frame = cap.read()

    # Convert frame to grayscale
    gray_frame = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)

    # Convert grayscale frame to RGB format
    rgb_frame = cv2.cvtColor(gray_frame, cv2.COLOR_GRAY2RGB)

    # Detect faces in the frame
    faces = face_cascade.detectMultiScale(gray_frame, scaleFactor=1.1, minNeighbors=5, minSize=(30, 30))
    for (x, y, w, h) in faces:
        # Extract the face ROI (Region of Interest)
        face_roi = rgb_frame[y:y + h, x:x + w]

        # Perform emotion analysis on the face ROI
        result = DeepFace.analyze(face_roi, actions=['emotion'], enforce_detection=False)

        # Get the dominant emotion and its confidence
        dominant_emotion = result[0]['dominant_emotion']
        confidence = result[0]['emotion'][dominant_emotion]

        # Only display the emotion if confidence is above the threshold
        if confidence > confidence_threshold:
            # Draw rectangle around face and label with predicted emotion
            cv2.rectangle(frame, (x, y), (x + w, y + h), (0, 0, 255), 2)
            cv2.putText(frame, dominant_emotion, (x, y - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.9, (0, 0, 255), 2)

    # Display the resulting frame
    cv2.imshow('Real-time Emotion Detection', frame)

    # Press 'q' to exit
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break

# Release the capture and close all windows
cap.release()
cv2.destroyAllWindows()

在此代码中：

1. 我们设置了一个 confidence_threshold 来定义显示情绪所需的最小置信度级别。
2. 我们从 DeepFace.analyze() 的结果中提取了 dominant_emotion 和它的 confidence 。
3. 我们仅在 confidence 高于 confidence_threshold 时才显示检测到的情绪。

通过调整 confidence_threshold ，可以控制在图像上绘制情绪标签之前的置信度级别。更高的阈值（例如，0.8 或 0.9）将导致更少但可能更准确的情绪检测，而较低的阈值将显示更多检测结果，但可能包括更多误报。