2020-08-20 02:05:34 +02:00
|
|
|
"use strict";
|
|
|
|
Object.defineProperty(exports, "__esModule", { value: true });
|
|
|
|
exports.FaceProcessor = void 0;
|
|
|
|
const tf = require("@tensorflow/tfjs-core");
|
|
|
|
const fullyConnectedLayer_1 = require("../common/fullyConnectedLayer");
|
|
|
|
const dom_1 = require("../dom");
|
|
|
|
const NeuralNetwork_1 = require("../NeuralNetwork");
|
|
|
|
const extractParams_1 = require("./extractParams");
|
|
|
|
const extractParamsFromWeigthMap_1 = require("./extractParamsFromWeigthMap");
|
|
|
|
const util_1 = require("./util");
|
|
|
|
class FaceProcessor extends NeuralNetwork_1.NeuralNetwork {
|
2020-08-18 14:04:33 +02:00
|
|
|
constructor(_name, faceFeatureExtractor) {
|
|
|
|
super(_name);
|
|
|
|
this._faceFeatureExtractor = faceFeatureExtractor;
|
|
|
|
}
|
|
|
|
get faceFeatureExtractor() {
|
|
|
|
return this._faceFeatureExtractor;
|
|
|
|
}
|
|
|
|
runNet(input) {
|
|
|
|
const { params } = this;
|
|
|
|
if (!params) {
|
|
|
|
throw new Error(`${this._name} - load model before inference`);
|
|
|
|
}
|
|
|
|
return tf.tidy(() => {
|
2020-08-20 02:05:34 +02:00
|
|
|
const bottleneckFeatures = input instanceof dom_1.NetInput
|
2020-08-18 14:04:33 +02:00
|
|
|
? this.faceFeatureExtractor.forwardInput(input)
|
|
|
|
: input;
|
2020-08-20 02:05:34 +02:00
|
|
|
return fullyConnectedLayer_1.fullyConnectedLayer(bottleneckFeatures.as2D(bottleneckFeatures.shape[0], -1), params.fc);
|
2020-08-18 14:04:33 +02:00
|
|
|
});
|
|
|
|
}
|
|
|
|
dispose(throwOnRedispose = true) {
|
|
|
|
this.faceFeatureExtractor.dispose(throwOnRedispose);
|
|
|
|
super.dispose(throwOnRedispose);
|
|
|
|
}
|
|
|
|
loadClassifierParams(weights) {
|
|
|
|
const { params, paramMappings } = this.extractClassifierParams(weights);
|
|
|
|
this._params = params;
|
|
|
|
this._paramMappings = paramMappings;
|
|
|
|
}
|
|
|
|
extractClassifierParams(weights) {
|
2020-08-20 02:05:34 +02:00
|
|
|
return extractParams_1.extractParams(weights, this.getClassifierChannelsIn(), this.getClassifierChannelsOut());
|
2020-08-18 14:04:33 +02:00
|
|
|
}
|
|
|
|
extractParamsFromWeigthMap(weightMap) {
|
2020-08-20 02:05:34 +02:00
|
|
|
const { featureExtractorMap, classifierMap } = util_1.seperateWeightMaps(weightMap);
|
2020-08-18 14:04:33 +02:00
|
|
|
this.faceFeatureExtractor.loadFromWeightMap(featureExtractorMap);
|
2020-08-20 02:05:34 +02:00
|
|
|
return extractParamsFromWeigthMap_1.extractParamsFromWeigthMap(classifierMap);
|
2020-08-18 14:04:33 +02:00
|
|
|
}
|
|
|
|
extractParams(weights) {
|
|
|
|
const cIn = this.getClassifierChannelsIn();
|
|
|
|
const cOut = this.getClassifierChannelsOut();
|
|
|
|
const classifierWeightSize = (cOut * cIn) + cOut;
|
|
|
|
const featureExtractorWeights = weights.slice(0, weights.length - classifierWeightSize);
|
|
|
|
const classifierWeights = weights.slice(weights.length - classifierWeightSize);
|
|
|
|
this.faceFeatureExtractor.extractWeights(featureExtractorWeights);
|
|
|
|
return this.extractClassifierParams(classifierWeights);
|
|
|
|
}
|
|
|
|
}
|
2020-08-20 02:05:34 +02:00
|
|
|
exports.FaceProcessor = FaceProcessor;
|
2020-08-18 14:04:33 +02:00
|
|
|
//# sourceMappingURL=FaceProcessor.js.map
|