@tensorflow/tfjs-core
Version:
Hardware-accelerated JavaScript library for machine intelligence
368 lines • 19.2 kB
JavaScript
"use strict";
var __awaiter = (this && this.__awaiter) || function (thisArg, _arguments, P, generator) {
return new (P || (P = Promise))(function (resolve, reject) {
function fulfilled(value) { try { step(generator.next(value)); } catch (e) { reject(e); } }
function rejected(value) { try { step(generator["throw"](value)); } catch (e) { reject(e); } }
function step(result) { result.done ? resolve(result.value) : new P(function (resolve) { resolve(result.value); }).then(fulfilled, rejected); }
step((generator = generator.apply(thisArg, _arguments || [])).next());
});
};
var __generator = (this && this.__generator) || function (thisArg, body) {
var _ = { label: 0, sent: function() { if (t[0] & 1) throw t[1]; return t[1]; }, trys: [], ops: [] }, f, y, t, g;
return g = { next: verb(0), "throw": verb(1), "return": verb(2) }, typeof Symbol === "function" && (g[Symbol.iterator] = function() { return this; }), g;
function verb(n) { return function (v) { return step([n, v]); }; }
function step(op) {
if (f) throw new TypeError("Generator is already executing.");
while (_) try {
if (f = 1, y && (t = op[0] & 2 ? y["return"] : op[0] ? y["throw"] || ((t = y["return"]) && t.call(y), 0) : y.next) && !(t = t.call(y, op[1])).done) return t;
if (y = 0, t) op = [op[0] & 2, t.value];
switch (op[0]) {
case 0: case 1: t = op; break;
case 4: _.label++; return { value: op[1], done: false };
case 5: _.label++; y = op[1]; op = [0]; continue;
case 7: op = _.ops.pop(); _.trys.pop(); continue;
default:
if (!(t = _.trys, t = t.length > 0 && t[t.length - 1]) && (op[0] === 6 || op[0] === 2)) { _ = 0; continue; }
if (op[0] === 3 && (!t || (op[1] > t[0] && op[1] < t[3]))) { _.label = op[1]; break; }
if (op[0] === 6 && _.label < t[1]) { _.label = t[1]; t = op; break; }
if (t && _.label < t[2]) { _.label = t[2]; _.ops.push(op); break; }
if (t[2]) _.ops.pop();
_.trys.pop(); continue;
}
op = body.call(thisArg, _);
} catch (e) { op = [6, e]; y = 0; } finally { f = t = 0; }
if (op[0] & 5) throw op[1]; return { value: op[0] ? op[1] : void 0, done: true };
}
};
var _this = this;
Object.defineProperty(exports, "__esModule", { value: true });
var tf = require("../index");
var jasmine_util_1 = require("../jasmine_util");
var test_util_1 = require("../test_util");
jasmine_util_1.describeWithFlags('nonMaxSuppression', test_util_1.ALL_ENVS, function () {
it('select from three clusters', function () {
var boxes = tf.tensor2d([
0, 0, 1, 1, 0, 0.1, 1, 1.1, 0, -0.1, 1, 0.9,
0, 10, 1, 11, 0, 10.1, 1, 11.1, 0, 100, 1, 101
], [6, 4]);
var scores = tf.tensor1d([0.9, 0.75, 0.6, 0.95, 0.5, 0.3]);
var maxOutputSize = 3;
var iouThreshold = 0.5;
var scoreThreshold = 0;
var indices = tf.image.nonMaxSuppression(boxes, scores, maxOutputSize, iouThreshold, scoreThreshold);
expect(indices.shape).toEqual([3]);
test_util_1.expectArraysEqual(indices, [3, 0, 5]);
});
it('select from three clusters flipped coordinates', function () {
var boxes = tf.tensor2d([
1, 1, 0, 0, 0, 0.1, 1, 1.1, 0, .9, 1, -0.1,
0, 10, 1, 11, 1, 10.1, 0, 11.1, 1, 101, 0, 100
], [6, 4]);
var scores = tf.tensor1d([0.9, 0.75, 0.6, 0.95, 0.5, 0.3]);
var maxOutputSize = 3;
var iouThreshold = 0.5;
var scoreThreshold = 0;
var indices = tf.image.nonMaxSuppression(boxes, scores, maxOutputSize, iouThreshold, scoreThreshold);
expect(indices.shape).toEqual([3]);
test_util_1.expectArraysEqual(indices, [3, 0, 5]);
});
it('select at most two boxes from three clusters', function () {
var boxes = tf.tensor2d([
0, 0, 1, 1, 0, 0.1, 1, 1.1, 0, -0.1, 1, 0.9,
0, 10, 1, 11, 0, 10.1, 1, 11.1, 0, 100, 1, 101
], [6, 4]);
var scores = tf.tensor1d([0.9, 0.75, 0.6, 0.95, 0.5, 0.3]);
var maxOutputSize = 2;
var iouThreshold = 0.5;
var scoreThreshold = 0;
var indices = tf.image.nonMaxSuppression(boxes, scores, maxOutputSize, iouThreshold, scoreThreshold);
expect(indices.shape).toEqual([2]);
test_util_1.expectArraysEqual(indices, [3, 0]);
});
it('select at most thirty boxes from three clusters', function () {
var boxes = tf.tensor2d([
0, 0, 1, 1, 0, 0.1, 1, 1.1, 0, -0.1, 1, 0.9,
0, 10, 1, 11, 0, 10.1, 1, 11.1, 0, 100, 1, 101
], [6, 4]);
var scores = tf.tensor1d([0.9, 0.75, 0.6, 0.95, 0.5, 0.3]);
var maxOutputSize = 30;
var iouThreshold = 0.5;
var scoreThreshold = 0;
var indices = tf.image.nonMaxSuppression(boxes, scores, maxOutputSize, iouThreshold, scoreThreshold);
expect(indices.shape).toEqual([3]);
test_util_1.expectArraysEqual(indices, [3, 0, 5]);
});
it('select single box', function () {
var boxes = tf.tensor2d([0, 0, 1, 1], [1, 4]);
var scores = tf.tensor1d([0.9]);
var maxOutputSize = 3;
var iouThreshold = 0.5;
var scoreThreshold = 0;
var indices = tf.image.nonMaxSuppression(boxes, scores, maxOutputSize, iouThreshold, scoreThreshold);
expect(indices.shape).toEqual([1]);
test_util_1.expectArraysEqual(indices, [0]);
});
it('select from ten identical boxes', function () {
var boxes = tf.tensor2d([0, 0, 1, 1], [1, 4]);
var scores = tf.tensor1d([0.9]);
var maxOutputSize = 3;
var iouThreshold = 0.5;
var scoreThreshold = 0;
var indices = tf.image.nonMaxSuppression(boxes, scores, maxOutputSize, iouThreshold, scoreThreshold);
expect(indices.shape).toEqual([1]);
test_util_1.expectArraysEqual(indices, [0]);
});
it('select from ten identical boxes', function () {
var numBoxes = 10;
var corners = new Array(numBoxes)
.fill(0)
.map(function (_) { return [0, 0, 1, 1]; })
.reduce(function (arr, curr) { return arr.concat(curr); });
var boxes = tf.tensor2d(corners, [numBoxes, 4]);
var scores = tf.tensor1d(Array(numBoxes).fill(0.9));
var maxOutputSize = 3;
var iouThreshold = 0.5;
var scoreThreshold = 0;
var indices = tf.image.nonMaxSuppression(boxes, scores, maxOutputSize, iouThreshold, scoreThreshold);
expect(indices.shape).toEqual([1]);
test_util_1.expectArraysEqual(indices, [0]);
});
it('inconsistent box and score shapes', function () {
var boxes = tf.tensor2d([
0, 0, 1, 1, 0, 0.1, 1, 1.1, 0, -0.1, 1, 0.9,
0, 10, 1, 11, 0, 10.1, 1, 11.1, 0, 100, 1, 101
], [6, 4]);
var scores = tf.tensor1d([0.9, 0.75, 0.6, 0.95, 0.5]);
var maxOutputSize = 30;
var iouThreshold = 0.5;
var scoreThreshold = 0;
expect(function () { return tf.image.nonMaxSuppression(boxes, scores, maxOutputSize, iouThreshold, scoreThreshold); })
.toThrowError(/scores has incompatible shape with boxes/);
});
it('invalid iou threshold', function () {
var boxes = tf.tensor2d([0, 0, 1, 1], [1, 4]);
var scores = tf.tensor1d([0.9]);
var maxOutputSize = 3;
var iouThreshold = 1.2;
var scoreThreshold = 0;
expect(function () { return tf.image.nonMaxSuppression(boxes, scores, maxOutputSize, iouThreshold, scoreThreshold); })
.toThrowError(/iouThreshold must be in \[0, 1\]/);
});
it('empty input', function () {
var boxes = tf.tensor2d([], [0, 4]);
var scores = tf.tensor1d([]);
var maxOutputSize = 3;
var iouThreshold = 0.5;
var scoreThreshold = 0;
var indices = tf.image.nonMaxSuppression(boxes, scores, maxOutputSize, iouThreshold, scoreThreshold);
expect(indices.shape).toEqual([0]);
test_util_1.expectArraysEqual(indices, []);
});
it('accepts a tensor-like object', function () {
var boxes = [[0, 0, 1, 1], [0, 1, 1, 2]];
var scores = [1, 2];
var indices = tf.image.nonMaxSuppression(boxes, scores, 10);
expect(indices.shape).toEqual([2]);
expect(indices.dtype).toEqual('int32');
test_util_1.expectArraysEqual(indices, [1, 0]);
});
});
jasmine_util_1.describeWithFlags('nonMaxSuppressionAsync', test_util_1.ALL_ENVS, function () {
it('select from three clusters', function () { return __awaiter(_this, void 0, void 0, function () {
var boxes, scores, maxOutputSize, iouThreshold, scoreThreshold, indices;
return __generator(this, function (_a) {
switch (_a.label) {
case 0:
boxes = tf.tensor2d([
0, 0, 1, 1, 0, 0.1, 1, 1.1, 0, -0.1, 1, 0.9,
0, 10, 1, 11, 0, 10.1, 1, 11.1, 0, 100, 1, 101
], [6, 4]);
scores = tf.tensor1d([0.9, 0.75, 0.6, 0.95, 0.5, 0.3]);
maxOutputSize = 3;
iouThreshold = 0.5;
scoreThreshold = 0;
return [4, tf.image.nonMaxSuppressionAsync(boxes, scores, maxOutputSize, iouThreshold, scoreThreshold)];
case 1:
indices = _a.sent();
expect(indices.shape).toEqual([3]);
test_util_1.expectArraysEqual(indices, [3, 0, 5]);
return [2];
}
});
}); });
it('accepts a tensor-like object', function () { return __awaiter(_this, void 0, void 0, function () {
var boxes, scores, indices;
return __generator(this, function (_a) {
switch (_a.label) {
case 0:
boxes = [[0, 0, 1, 1], [0, 1, 1, 2]];
scores = [1, 2];
return [4, tf.image.nonMaxSuppressionAsync(boxes, scores, 10)];
case 1:
indices = _a.sent();
expect(indices.shape).toEqual([2]);
expect(indices.dtype).toEqual('int32');
test_util_1.expectArraysEqual(indices, [1, 0]);
return [2];
}
});
}); });
});
jasmine_util_1.describeWithFlags('cropAndResize', test_util_1.ALL_ENVS, function () {
it('1x1-bilinear', function () {
var image = tf.tensor4d([1, 2, 3, 4], [1, 2, 2, 1]);
var boxes = tf.tensor2d([0, 0, 1, 1], [1, 4]);
var boxInd = tf.tensor1d([0], 'int32');
var output = tf.image.cropAndResize(image, boxes, boxInd, [1, 1], 'bilinear', 0);
expect(output.shape).toEqual([1, 1, 1, 1]);
test_util_1.expectArraysClose(output, [2.5]);
});
it('1x1-nearest', function () {
var image = tf.tensor4d([1, 2, 3, 4], [1, 2, 2, 1]);
var boxes = tf.tensor2d([0, 0, 1, 1], [1, 4]);
var boxInd = tf.tensor1d([0], 'int32');
var output = tf.image.cropAndResize(image, boxes, boxInd, [1, 1], 'nearest', 0);
expect(output.shape).toEqual([1, 1, 1, 1]);
test_util_1.expectArraysClose(output, [4.0]);
});
it('1x1Flipped-bilinear', function () {
var image = tf.tensor4d([1, 2, 3, 4], [1, 2, 2, 1]);
var boxes = tf.tensor2d([1, 1, 0, 0], [1, 4]);
var boxInd = tf.tensor1d([0], 'int32');
var output = tf.image.cropAndResize(image, boxes, boxInd, [1, 1], 'bilinear', 0);
expect(output.shape).toEqual([1, 1, 1, 1]);
test_util_1.expectArraysClose(output, [2.5]);
});
it('1x1Flipped-nearest', function () {
var image = tf.tensor4d([1, 2, 3, 4], [1, 2, 2, 1]);
var boxes = tf.tensor2d([1, 1, 0, 0], [1, 4]);
var boxInd = tf.tensor1d([0], 'int32');
var output = tf.image.cropAndResize(image, boxes, boxInd, [1, 1], 'nearest', 0);
expect(output.shape).toEqual([1, 1, 1, 1]);
test_util_1.expectArraysClose(output, [4.0]);
});
it('3x3-bilinear', function () {
var image = tf.tensor4d([1, 2, 3, 4], [1, 2, 2, 1]);
var boxes = tf.tensor2d([0, 0, 1, 1], [1, 4]);
var boxInd = tf.tensor1d([0], 'int32');
var output = tf.image.cropAndResize(image, boxes, boxInd, [3, 3], 'bilinear', 0);
expect(output.shape).toEqual([1, 3, 3, 1]);
test_util_1.expectArraysClose(output, [1, 1.5, 2, 2, 2.5, 3, 3, 3.5, 4]);
});
it('3x3-nearest', function () {
var image = tf.tensor4d([1, 2, 3, 4], [1, 2, 2, 1]);
var boxes = tf.tensor2d([0, 0, 1, 1], [1, 4]);
var boxInd = tf.tensor1d([0], 'int32');
var output = tf.image.cropAndResize(image, boxes, boxInd, [3, 3], 'nearest', 0);
expect(output.shape).toEqual([1, 3, 3, 1]);
test_util_1.expectArraysClose(output, [1, 2, 2, 3, 4, 4, 3, 4, 4]);
});
it('3x3Flipped-bilinear', function () {
var image = tf.tensor4d([1, 2, 3, 4], [1, 2, 2, 1]);
var boxes = tf.tensor2d([1, 1, 0, 0], [1, 4]);
var boxInd = tf.tensor1d([0], 'int32');
var output = tf.image.cropAndResize(image, boxes, boxInd, [3, 3], 'bilinear', 0);
expect(output.shape).toEqual([1, 3, 3, 1]);
test_util_1.expectArraysClose(output, [4, 3.5, 3, 3, 2.5, 2, 2, 1.5, 1]);
});
it('3x3Flipped-nearest', function () {
var image = tf.tensor4d([1, 2, 3, 4], [1, 2, 2, 1]);
var boxes = tf.tensor2d([1, 1, 0, 0], [1, 4]);
var boxInd = tf.tensor1d([0], 'int32');
var output = tf.image.cropAndResize(image, boxes, boxInd, [3, 3], 'nearest', 0);
expect(output.shape).toEqual([1, 3, 3, 1]);
test_util_1.expectArraysClose(output, [4, 4, 3, 4, 4, 3, 2, 2, 1]);
});
it('3x3to2x2-bilinear', function () {
var image = tf.tensor4d([1, 2, 3, 4, 5, 6, 7, 8, 9], [1, 3, 3, 1]);
var boxes = tf.tensor2d([0, 0, 1, 1, 0, 0, 0.5, 0.5], [2, 4]);
var boxInd = tf.tensor1d([0, 0], 'int32');
var output = tf.image.cropAndResize(image, boxes, boxInd, [2, 2], 'bilinear', 0);
expect(output.shape).toEqual([2, 2, 2, 1]);
test_util_1.expectArraysClose(output, [1, 3, 7, 9, 1, 2, 4, 5]);
});
it('3x3to2x2-nearest', function () {
var image = tf.tensor4d([1, 2, 3, 4, 5, 6, 7, 8, 9], [1, 3, 3, 1]);
var boxes = tf.tensor2d([0, 0, 1, 1, 0, 0, 0.5, 0.5], [2, 4]);
var boxInd = tf.tensor1d([0, 0], 'int32');
var output = tf.image.cropAndResize(image, boxes, boxInd, [2, 2], 'nearest', 0);
expect(output.shape).toEqual([2, 2, 2, 1]);
test_util_1.expectArraysClose(output, [1, 3, 7, 9, 1, 2, 4, 5]);
});
it('3x3to2x2Flipped-bilinear', function () {
var image = tf.tensor4d([1, 2, 3, 4, 5, 6, 7, 8, 9], [1, 3, 3, 1]);
var boxes = tf.tensor2d([1, 1, 0, 0, 0.5, 0.5, 0, 0], [2, 4]);
var boxInd = tf.tensor1d([0, 0], 'int32');
var output = tf.image.cropAndResize(image, boxes, boxInd, [2, 2], 'bilinear', 0);
expect(output.shape).toEqual([2, 2, 2, 1]);
test_util_1.expectArraysClose(output, [9, 7, 3, 1, 5, 4, 2, 1]);
});
it('3x3to2x2Flipped-nearest', function () {
var image = tf.tensor4d([1, 2, 3, 4, 5, 6, 7, 8, 9], [1, 3, 3, 1]);
var boxes = tf.tensor2d([1, 1, 0, 0, 0.5, 0.5, 0, 0], [2, 4]);
var boxInd = tf.tensor1d([0, 0], 'int32');
var output = tf.image.cropAndResize(image, boxes, boxInd, [2, 2], 'nearest', 0);
expect(output.shape).toEqual([2, 2, 2, 1]);
test_util_1.expectArraysClose(output, [9, 7, 3, 1, 5, 4, 2, 1]);
});
it('3x3-BoxisRectangular', function () {
var image = tf.tensor4d([1, 2, 3, 4], [1, 2, 2, 1]);
var boxes = tf.tensor2d([0, 0, 1, 1.5], [1, 4]);
var boxInd = tf.tensor1d([0], 'int32');
var output = tf.image.cropAndResize(image, boxes, boxInd, [3, 3], 'bilinear', 0);
expect(output.shape).toEqual([1, 3, 3, 1]);
test_util_1.expectArraysClose(output, [1, 1.75, 0, 2, 2.75, 0, 3, 3.75, 0]);
});
it('3x3-BoxisRectangular-nearest', function () {
var image = tf.tensor4d([1, 2, 3, 4], [1, 2, 2, 1]);
var boxes = tf.tensor2d([0, 0, 1, 1.5], [1, 4]);
var boxInd = tf.tensor1d([0], 'int32');
var output = tf.image.cropAndResize(image, boxes, boxInd, [3, 3], 'nearest', 0);
expect(output.shape).toEqual([1, 3, 3, 1]);
test_util_1.expectArraysClose(output, [1, 2, 0, 3, 4, 0, 3, 4, 0]);
});
it('2x2to3x3-Extrapolated', function () {
var val = -1;
var image = tf.tensor4d([1, 2, 3, 4], [1, 2, 2, 1]);
var boxes = tf.tensor2d([-1, -1, 1, 1], [1, 4]);
var boxInd = tf.tensor1d([0], 'int32');
var output = tf.image.cropAndResize(image, boxes, boxInd, [3, 3], 'bilinear', val);
expect(output.shape).toEqual([1, 3, 3, 1]);
test_util_1.expectArraysClose(output, [val, val, val, val, 1, 2, val, 3, 4]);
});
it('2x2to3x3-Extrapolated-Float', function () {
var val = -1.5;
var image = tf.tensor4d([1, 2, 3, 4], [1, 2, 2, 1]);
var boxes = tf.tensor2d([-1, -1, 1, 1], [1, 4]);
var boxInd = tf.tensor1d([0], 'int32');
var output = tf.image.cropAndResize(image, boxes, boxInd, [3, 3], 'bilinear', val);
expect(output.shape).toEqual([1, 3, 3, 1]);
test_util_1.expectArraysClose(output, [val, val, val, val, 1, 2, val, 3, 4]);
});
it('2x2to3x3-NoCrop', function () {
var val = -1.0;
var image = tf.tensor4d([1, 2, 3, 4], [1, 2, 2, 1]);
var boxes = tf.tensor2d([], [0, 4]);
var boxInd = tf.tensor1d([], 'int32');
var output = tf.image.cropAndResize(image, boxes, boxInd, [3, 3], 'bilinear', val);
expect(output.shape).toEqual([0, 3, 3, 1]);
test_util_1.expectArraysClose(output, []);
});
it('MultipleBoxes-DifferentBoxes', function () {
var image = tf.tensor4d([1, 2, 3, 4, 5, 6, 7, 8], [2, 2, 2, 1]);
var boxes = tf.tensor2d([0, 0, 1, 1.5, 0, 0, 1.5, 1], [2, 4]);
var boxInd = tf.tensor1d([0, 1], 'int32');
var output = tf.image.cropAndResize(image, boxes, boxInd, [3, 3], 'bilinear', 0);
expect(output.shape).toEqual([2, 3, 3, 1]);
test_util_1.expectArraysClose(output, [1, 1.75, 0, 2, 2.75, 0, 3, 3.75, 0, 5, 5.5, 6, 6.5, 7, 7.5, 0, 0, 0]);
});
it('MultipleBoxes-DifferentBoxes-Nearest', function () {
var image = tf.tensor4d([1, 2, 3, 4, 5, 6, 7, 8], [2, 2, 2, 1]);
var boxes = tf.tensor2d([0, 0, 1, 1.5, 0, 0, 2, 1], [2, 4]);
var boxInd = tf.tensor1d([0, 1], 'int32');
var output = tf.image.cropAndResize(image, boxes, boxInd, [3, 3], 'nearest', 0);
expect(output.shape).toEqual([2, 3, 3, 1]);
test_util_1.expectArraysClose(output, [1, 2, 0, 3, 4, 0, 3, 4, 0, 5, 6, 6, 7, 8, 8, 0, 0, 0]);
});
});
//# sourceMappingURL=image_ops_test.js.map