找到需要識別的臉
faceDetector = vision.CascadeObjectDetector();

讀取一段錄影並從中獲取需識別的臉
videoFileReader = vision.VideoFileReader(‘tilted_face.avi’);
videoFrame = step(videoFileReader);
bbox = step(faceDetector, videoFrame);

在被檢測的臉周圍構建框架
videoFrame = insertShape(videoFrame, ‘Rectangle’, bbox);
figure; imshow(videoFrame); title(‘Detected face’);

將框架轉化為一系列的4個點，即使被識別的臉發生旋轉也能夠被看到
bboxPoints = bbox2points(bbox(1, :));

在臉部區域識別特徵點
points = detectMinEigenFeatures(rgb2gray(videoFrame), ‘ROI’, bbox);
（指定被保留特徵值的最小值,一般為1 ）

顯示特徵點
figure, imshow(videoFrame), hold on, title(‘Detected features’);
plot(points);

建立一個點的蹤跡將錯誤進行雙向限制，以使它即使在噪聲存在時也能正常表示
pointTracker = vision.PointTracker(‘MaxBidirectionalError’, 2);

用初始化框架的點的位置初始化蹤跡
points = points.Location;
initialize(pointTracker, points, videoFrame);

初始化一個視訊播放器顯示結果
videoPlayer = vision.VideoPlayer(‘Position’,…
[100 100 [size(videoFrame, 2), size(videoFrame, 1)]+30]);

複製點用於之前的點和現在的框架之間進行幾何轉換
oldPoints = points;
while ~isDone(videoFileReader)

建立下一個框架
videoFrame = step(videoFileReader);

跟隨點的蹤跡，注意有幾個點可能會丟失
[points, isFound] = step(pointTracker, videoFrame);
visiblePoints = points(isFound, :);
oldInliers = oldPoints(isFound, :);
if size(visiblePoints, 1) >= 2 % need at least 2 points

在之前的點和新的點之間進行幾何轉換並建立邊界線
[xform, oldInliers, visiblePoints] = estimateGeometricTransform(…
oldInliers, visiblePoints, ‘similarity’, ‘MaxDistance’, 4);

執行邊界點的轉換
bboxPoints = transformPointsForward(xform, bboxPoints);

在被追隨的物體周圍插入邊界框
bboxPolygon = reshape(bboxPoints’, 1, []);
videoFrame = insertShape(videoFrame, ‘Polygon’, bboxPolygon, …
‘LineWidth’, 2);
（reshape表示重新調整矩陣的行數、列數、維數）

顯示被追隨的點
videoFrame = insertMarker(videoFrame, visiblePoints, ‘+’, …
‘Color’, ‘white’); （在影象或視訊中插入標記）

重置點
oldPoints = visiblePoints;
setPoints(pointTracker, oldPoints); （設定跟隨點，用於點需要重新檢測且很多點丟失在跟蹤過程中，可以使用此方法。）
end

用視訊播放器顯示被註釋的視訊框架
step(videoPlayer, videoFrame);
end

清空
release(videoFileReader);
release(videoPlayer);
release(pointTracker);