Python | OpenCV 專案：即時人臉辨識系統

# 1_collect.py
import cv2
import os

# ===== 設定 =====
person_id      = 0                       # 每位人員指定一個唯一 ID（從 0 開始遞增）
person_name    = "PersonA"               # 人員姓名
video_path     = "assets/person_a.mp4"   # 來源影片路徑
sample_count   = 100                     # 目標蒐集樣本數
frame_interval = 3                       # 每 N 幀處理一次，避免連續幀過於相似
# ================

save_dir = f"dataset/{person_id}_{person_name}"
os.makedirs(save_dir, exist_ok=True)

face_cascade = cv2.CascadeClassifier(
    cv2.data.haarcascades + "haarcascade_frontalface_default.xml"
)

cap = cv2.VideoCapture(video_path)
if not cap.isOpened():
    print(f"無法開啟影片：{video_path}")
    exit()

count     = 0
frame_idx = 0
print(f"開始從影片蒐集 {person_name} 的人臉樣本，目標 {sample_count} 張")

while count < sample_count:
    ret, frame = cap.read()
    if not ret:
        print("影片播放完畢")
        break

    frame_idx += 1
    if frame_idx % frame_interval != 0:
        continue

    gray  = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
    faces = face_cascade.detectMultiScale(gray, 1.3, 5)

    for (x, y, w, h) in faces:
        face_img = gray[y:y+h, x:x+w]
        face_rsz = cv2.resize(face_img, (200, 200))
        filename = os.path.join(save_dir, f"{count:04d}.jpg")
        cv2.imwrite(filename, face_rsz)
        count += 1

        cv2.rectangle(frame, (x, y), (x+w, y+h), (0, 255, 0), 2)
        cv2.putText(frame, f"{count}/{sample_count}", (x, y - 10),
                    cv2.FONT_HERSHEY_SIMPLEX, 0.7, (0, 255, 0), 2)

    cv2.imshow("Collecting", frame)
    if cv2.waitKey(30) & 0xFF == ord("q"):
        break

cap.release()
cv2.destroyAllWindows()
print(f"完成！共蒐集 {count} 張樣本，儲存於 {save_dir}")

# 2_train.py
import cv2
import numpy as np
import os
import json

dataset_dir = "dataset"
faces       = []
labels      = []
label_map   = {}    # id → name

for folder in sorted(os.listdir(dataset_dir)):
    folder_path = os.path.join(dataset_dir, folder)
    if not os.path.isdir(folder_path):
        continue

    # 資料夾命名規則：{id}_{name}
    parts       = folder.split("_", 1)
    person_id   = int(parts[0])
    person_name = parts[1] if len(parts) > 1 else folder
    label_map[person_id] = person_name

    for fname in os.listdir(folder_path):
        if not fname.lower().endswith(".jpg"):
            continue
        img_path = os.path.join(folder_path, fname)
        img      = cv2.imread(img_path, cv2.IMREAD_GRAYSCALE)
        if img is not None:
            faces.append(img)
            labels.append(person_id)

print(f"載入完成：{len(faces)} 張樣本，{len(label_map)} 位人員")
print(f"人員清單：{label_map}")

recognizer = cv2.face.LBPHFaceRecognizer_create()
recognizer.train(faces, np.array(labels))

os.makedirs("models", exist_ok=True)
recognizer.save("models/face_model.yml")

with open("models/label_map.json", "w", encoding="utf-8") as f:
    json.dump({str(k): v for k, v in label_map.items()}, f, ensure_ascii=False)

print("訓練完成！模型已儲存至 model/face_model.yml")

# 3_recognize.py
import cv2
import json

# 載入模型與標籤
recognizer = cv2.face.LBPHFaceRecognizer_create()
recognizer.read("models/face_model.yml")

with open("models/label_map.json", encoding="utf-8") as f:
    raw_map   = json.load(f)
    label_map = {int(k): v for k, v in raw_map.items()}

face_cascade = cv2.CascadeClassifier(
    cv2.data.haarcascades + "haarcascade_frontalface_default.xml"
)

CONFIDENCE_THRESHOLD = 80
video_path = "assets/test.mp4"

cap = cv2.VideoCapture(video_path)
if not cap.isOpened():
    print(f"無法開啟影片：{video_path}")
    exit()
print("影片辨識啟動，按 q 離開")

while True:
    ret, frame = cap.read()
    if not ret:
        print("影片播放完畢")
        break

    gray  = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
    faces = face_cascade.detectMultiScale(gray, 1.3, 5)

    for (x, y, w, h) in faces:
        face_img = gray[y:y+h, x:x+w]
        face_rsz = cv2.resize(face_img, (200, 200))

        label, confidence = recognizer.predict(face_rsz)

        # LBPH：confidence 越低代表越相似（0 = 完全一致）
        if confidence < CONFIDENCE_THRESHOLD:
            name  = label_map.get(label, "Unknown")
            color = (0, 255, 0)
            text  = f"{name} ({confidence:.1f})"
        else:
            name  = "Unknown"
            color = (0, 0, 255)
            text  = f"Unknown ({confidence:.1f})"

        cv2.rectangle(frame, (x, y), (x+w, y+h), color, 2)
        cv2.putText(frame, text, (x, y - 10),
                    cv2.FONT_HERSHEY_SIMPLEX, 0.7, color, 2)

    cv2.imshow("Face Recognition", frame)
    if cv2.waitKey(30) & 0xFF == ord("q"):
        break

cap.release()
cv2.destroyAllWindows()