Implementation:Tensorflow Tfjs Resizing Layer

Knowledge Sources	Tensorflow_Tfjs TensorFlow.js API
Domains	Deep_Learning, Layers_API, Preprocessing
Last Updated	2026-02-10 06:00 GMT

Overview

The Resizing layer is an image preprocessing layer that resizes images to a fixed target height and width. It supports two interpolation methods: bilinear (default) and nearest. An optional cropToAspectRatio flag controls whether the image is cropped to preserve its aspect ratio or stretched to fill the target size.

Code Reference

Source Location

tfjs-layers/src/layers/preprocessing/image_resizing.ts (GitHub)

Key Imports

import {image, Rank, serialization, Tensor, tidy} from '@tensorflow/tfjs-core';
import {Layer, LayerArgs} from '../../engine/topology';
import {ValueError} from '../../errors';

Layer Class

export class Resizing extends Layer {
  static className = 'Resizing';
  constructor(args: ResizingArgs);
  override computeOutputShape(inputShape: Shape | Shape[]): Shape | Shape[];
  override getConfig(): serialization.ConfigDict;
  override call(inputs: Tensor<Rank.R3> | Tensor<Rank.R4>, kwargs: Kwargs): Tensor[] | Tensor;
}

ResizingArgs

export interface ResizingArgs extends LayerArgs {
  height: number;                      // target height
  width: number;                       // target width
  interpolation?: InterpolationType;   // 'bilinear' (default) or 'nearest'
  cropToAspectRatio?: boolean;         // default: false
}

Implementation Details

• Validates the interpolation method against supported options (bilinear, nearest).
• Uses image.resizeBilinear or image.resizeNearestNeighbor from tfjs-core.
• The alignCorners parameter is set to !cropToAspectRatio.

I/O Contract

Method	Input	Output
call	3D or 4D image tensor	Resized tensor with spatial dims [height, width]
computeOutputShape	Input shape	[height, width, numChannels]

Usage Example

import * as tf from '@tensorflow/tfjs';

const resizer = tf.layers.resizing({
  height: 224,
  width: 224,
  interpolation: 'bilinear'
});

const img = tf.randomNormal([1, 480, 640, 3]);
const resized = resizer.apply(img);  // shape: [1, 224, 224, 3]

Related Pages

• Implementation:Tensorflow_Tfjs_CenterCrop_Layer - Center crop preprocessing layer
• Implementation:Tensorflow_Tfjs_Rescaling_Layer - Pixel value rescaling layer
• Implementation:Tensorflow_Tfjs_RandomHeight_Layer - Random height augmentation
• Implementation:Tensorflow_Tfjs_RandomWidth_Layer - Random width augmentation