chenyc
2025-05-29 92f69c57b920cf62ecc9f15f9ed196fa26dbf2ac
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
"use strict";
Object.defineProperty(exports, "__esModule", { value: true });
var tslib_1 = require("tslib");
var tf = require("@tensorflow/tfjs-core");
var dom_1 = require("../dom");
var FaceFeatureExtractor_1 = require("../faceFeatureExtractor/FaceFeatureExtractor");
var FaceProcessor_1 = require("../faceProcessor/FaceProcessor");
var FaceExpressions_1 = require("./FaceExpressions");
var FaceExpressionNet = /** @class */ (function (_super) {
    tslib_1.__extends(FaceExpressionNet, _super);
    function FaceExpressionNet(faceFeatureExtractor) {
        if (faceFeatureExtractor === void 0) { faceFeatureExtractor = new FaceFeatureExtractor_1.FaceFeatureExtractor(); }
        return _super.call(this, 'FaceExpressionNet', faceFeatureExtractor) || this;
    }
    FaceExpressionNet.prototype.forwardInput = function (input) {
        var _this = this;
        return tf.tidy(function () { return tf.softmax(_this.runNet(input)); });
    };
    FaceExpressionNet.prototype.forward = function (input) {
        return tslib_1.__awaiter(this, void 0, void 0, function () {
            var _a;
            return tslib_1.__generator(this, function (_b) {
                switch (_b.label) {
                    case 0:
                        _a = this.forwardInput;
                        return [4 /*yield*/, dom_1.toNetInput(input)];
                    case 1: return [2 /*return*/, _a.apply(this, [_b.sent()])];
                }
            });
        });
    };
    FaceExpressionNet.prototype.predictExpressions = function (input) {
        return tslib_1.__awaiter(this, void 0, void 0, function () {
            var netInput, out, probabilitesByBatch, predictionsByBatch;
            var _this = this;
            return tslib_1.__generator(this, function (_a) {
                switch (_a.label) {
                    case 0: return [4 /*yield*/, dom_1.toNetInput(input)];
                    case 1:
                        netInput = _a.sent();
                        return [4 /*yield*/, this.forwardInput(netInput)];
                    case 2:
                        out = _a.sent();
                        return [4 /*yield*/, Promise.all(tf.unstack(out).map(function (t) { return tslib_1.__awaiter(_this, void 0, void 0, function () {
                                var data;
                                return tslib_1.__generator(this, function (_a) {
                                    switch (_a.label) {
                                        case 0: return [4 /*yield*/, t.data()];
                                        case 1:
                                            data = _a.sent();
                                            t.dispose();
                                            return [2 /*return*/, data];
                                    }
                                });
                            }); }))];
                    case 3:
                        probabilitesByBatch = _a.sent();
                        out.dispose();
                        predictionsByBatch = probabilitesByBatch
                            .map(function (probabilites) { return new FaceExpressions_1.FaceExpressions(probabilites); });
                        return [2 /*return*/, netInput.isBatchInput
                                ? predictionsByBatch
                                : predictionsByBatch[0]];
                }
            });
        });
    };
    FaceExpressionNet.prototype.getDefaultModelName = function () {
        return 'face_expression_model';
    };
    FaceExpressionNet.prototype.getClassifierChannelsIn = function () {
        return 256;
    };
    FaceExpressionNet.prototype.getClassifierChannelsOut = function () {
        return 7;
    };
    return FaceExpressionNet;
}(FaceProcessor_1.FaceProcessor));
exports.FaceExpressionNet = FaceExpressionNet;
//# sourceMappingURL=FaceExpressionNet.js.map