|
| 1 | +// Copyright 2025 Google LLC |
| 2 | +// |
| 3 | +// Licensed under the Apache License, Version 2.0 (the "License"); |
| 4 | +// you may not use this file except in compliance with the License. |
| 5 | +// You may obtain a copy of the License at |
| 6 | +// |
| 7 | +// https://www.apache.org/licenses/LICENSE-2.0 |
| 8 | +// |
| 9 | +// Unless required by applicable law or agreed to in writing, software |
| 10 | +// distributed under the License is distributed on an "AS IS" BASIS, |
| 11 | +// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| 12 | +// See the License for the specific language governing permissions and |
| 13 | +// limitations under the License. |
| 14 | + |
| 15 | +'use strict'; |
| 16 | + |
| 17 | +// [START googlegenaisdk_boundingbox_with_txt_img] |
| 18 | +const {GoogleGenAI} = require('@google/genai'); |
| 19 | + |
| 20 | +const {createCanvas, loadImage} = require('canvas'); |
| 21 | +const fetch = require('node-fetch'); |
| 22 | +const fs = require('fs'); |
| 23 | + |
| 24 | +const GOOGLE_CLOUD_PROJECT = process.env.GOOGLE_CLOUD_PROJECT; |
| 25 | +const GOOGLE_CLOUD_LOCATION = process.env.GOOGLE_CLOUD_LOCATION || 'global'; |
| 26 | + |
| 27 | +async function fetchImageAsBase64(uri) { |
| 28 | + const response = await fetch(uri); |
| 29 | + const buffer = await response.buffer(); |
| 30 | + return buffer.toString('base64'); |
| 31 | +} |
| 32 | + |
| 33 | +async function plotBoundingBoxes(imageUri, boundingBoxes) { |
| 34 | + console.log('Creating bounding boxes'); |
| 35 | + const image = await loadImage(imageUri); |
| 36 | + const canvas = createCanvas(image.width, image.height); |
| 37 | + const ctx = canvas.getContext('2d'); |
| 38 | + |
| 39 | + ctx.drawImage(image, 0, 0); |
| 40 | + |
| 41 | + const colors = ['red', 'blue', 'green', 'orange']; |
| 42 | + |
| 43 | + boundingBoxes.forEach((bbox, i) => { |
| 44 | + const [yMin, xMin, yMax, xMax] = bbox.box_2d; |
| 45 | + |
| 46 | + const absYMin = Math.floor((yMin / 1000) * image.height); |
| 47 | + const absXMin = Math.floor((xMin / 1000) * image.width); |
| 48 | + const absYMax = Math.floor((yMax / 1000) * image.height); |
| 49 | + const absXMax = Math.floor((xMax / 1000) * image.width); |
| 50 | + |
| 51 | + ctx.strokeStyle = colors[i % colors.length]; |
| 52 | + ctx.lineWidth = 4; |
| 53 | + ctx.strokeRect(absXMin, absYMin, absXMax - absXMin, absYMax - absYMin); |
| 54 | + |
| 55 | + ctx.fillStyle = colors[i % colors.length]; |
| 56 | + ctx.font = '20px Arial'; |
| 57 | + ctx.fillText(bbox.label, absXMin + 8, absYMin + 20); |
| 58 | + }); |
| 59 | + |
| 60 | + fs.writeFileSync('output.png', canvas.toBuffer('image/png')); |
| 61 | + console.log('Saved output to file: output.png'); |
| 62 | +} |
| 63 | + |
| 64 | +async function createBoundingBox( |
| 65 | + projectId = GOOGLE_CLOUD_PROJECT, |
| 66 | + location = GOOGLE_CLOUD_LOCATION |
| 67 | +) { |
| 68 | + const client = new GoogleGenAI({ |
| 69 | + vertexai: true, |
| 70 | + project: projectId, |
| 71 | + location: location, |
| 72 | + }); |
| 73 | + |
| 74 | + const systemInstruction = ` |
| 75 | + Return bounding boxes as an array with labels. |
| 76 | + Never return masks. Limit to 25 objects. |
| 77 | + If an object is present multiple times, give each object a unique label |
| 78 | + according to its distinct characteristics (colors, size, position, etc). |
| 79 | + `; |
| 80 | + |
| 81 | + const safetySettings = [ |
| 82 | + { |
| 83 | + category: 'HARM_CATEGORY_DANGEROUS_CONTENT', |
| 84 | + threshold: 'BLOCK_ONLY_HIGH', |
| 85 | + }, |
| 86 | + ]; |
| 87 | + |
| 88 | + const imageUri = |
| 89 | + 'https://storage.googleapis.com/generativeai-downloads/images/socks.jpg'; |
| 90 | + const base64Image = await fetchImageAsBase64(imageUri); |
| 91 | + |
| 92 | + const boundingBoxSchema = { |
| 93 | + type: 'ARRAY', |
| 94 | + description: 'List of bounding boxes for detected objects', |
| 95 | + items: { |
| 96 | + type: 'OBJECT', |
| 97 | + title: 'BoundingBox', |
| 98 | + description: 'Represents a bounding box with coordinates and label', |
| 99 | + properties: { |
| 100 | + box_2d: { |
| 101 | + type: 'ARRAY', |
| 102 | + description: |
| 103 | + 'Bounding box coordinates in format [y_min, x_min, y_max, x_max]', |
| 104 | + items: { |
| 105 | + type: 'INTEGER', |
| 106 | + format: 'int32', |
| 107 | + }, |
| 108 | + minItems: 4, |
| 109 | + maxItems: 4, |
| 110 | + }, |
| 111 | + label: { |
| 112 | + type: 'STRING', |
| 113 | + description: 'Label describing the object within the bounding box', |
| 114 | + }, |
| 115 | + }, |
| 116 | + required: ['box_2d', 'label'], |
| 117 | + }, |
| 118 | + }; |
| 119 | + |
| 120 | + const response = await client.models.generateContent({ |
| 121 | + model: 'gemini-2.5-flash', |
| 122 | + contents: [ |
| 123 | + { |
| 124 | + role: 'user', |
| 125 | + parts: [ |
| 126 | + { |
| 127 | + text: 'Output the positions of the socks with a face. Label according to position in the image', |
| 128 | + }, |
| 129 | + { |
| 130 | + inlineData: { |
| 131 | + data: base64Image, |
| 132 | + mimeType: 'image/jpeg', |
| 133 | + }, |
| 134 | + }, |
| 135 | + ], |
| 136 | + }, |
| 137 | + ], |
| 138 | + config: { |
| 139 | + systemInstruction: systemInstruction, |
| 140 | + safetySettings: safetySettings, |
| 141 | + responseMimeType: 'application/json', |
| 142 | + temperature: 0.5, |
| 143 | + responseSchema: boundingBoxSchema, |
| 144 | + }, |
| 145 | + }); |
| 146 | + |
| 147 | + const candidate = response.candidates[0].content.parts[0].text; |
| 148 | + const boundingBoxes = JSON.parse(candidate); |
| 149 | + |
| 150 | + console.log('Bounding boxes:', boundingBoxes); |
| 151 | + |
| 152 | + await plotBoundingBoxes(imageUri, boundingBoxes); |
| 153 | + return boundingBoxes; |
| 154 | +} |
| 155 | +// [END googlegenaisdk_boundingbox_with_txt_img] |
| 156 | + |
| 157 | +module.exports = { |
| 158 | + createBoundingBox, |
| 159 | +}; |
0 commit comments