cn.smartjavaai.face.translator

类 YoloV5FaceTranslator

java.lang.Object

cn.smartjavaai.face.translator.YoloV5FaceTranslator

所有已实现的接口:

ai.djl.translate.PostProcessor<ai.djl.modality.cv.output.DetectedObjects>, ai.djl.translate.PreProcessor<ai.djl.modality.cv.Image>, ai.djl.translate.Translator<ai.djl.modality.cv.Image,ai.djl.modality.cv.output.DetectedObjects>

public class YoloV5FaceTranslator
extends Object
implements ai.djl.translate.Translator<ai.djl.modality.cv.Image,ai.djl.modality.cv.output.DetectedObjects>

YoloV5Translator

作者:

dwj

嵌套类概要

嵌套类

限定符和类型	类和说明
static class	YoloV5FaceTranslator.Builder
static class	YoloV5FaceTranslator.YoloOutputType

字段概要

字段

限定符和类型	字段和说明
protected boolean	applyRatio
protected List<String>	classes
protected int	height
protected ai.djl.translate.Pipeline	pipeline
protected boolean	removePadding
protected float	threshold
protected int	width

构造器概要

构造器

限定符	构造器和说明
protected	YoloV5FaceTranslator(YoloV5FaceTranslator.Builder builder) Constructs an ImageTranslator with the provided builder.

方法概要

限定符和类型	方法和说明
static YoloV5FaceTranslator.Builder	builder() Creates a builder to build a YoloV8Translator with specified arguments.
static YoloV5FaceTranslator.Builder	builder(Map<String,?> arguments) Creates a builder to build a YoloV8Translator with specified arguments.
protected ai.djl.modality.cv.output.DetectedObjects	nms(int imageWidth, int imageHeight, List<ai.djl.modality.cv.output.Landmark> boxes, List<Integer> classIds, List<Float> scores, float scale)
protected ai.djl.modality.cv.output.DetectedObjects	processFromBoxOutput(int imageWidth, int imageHeight, ai.djl.ndarray.NDList list, float scale)
ai.djl.ndarray.NDList	processInput(ai.djl.translate.TranslatorContext ctx, ai.djl.modality.cv.Image input)
ai.djl.modality.cv.output.DetectedObjects	processOutput(ai.djl.translate.TranslatorContext ctx, ai.djl.ndarray.NDList list)

从类继承的方法 java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

从接口继承的方法 ai.djl.translate.Translator

batchProcessInput, batchProcessOutput, getBatchifier, getExpansions, prepare

字段详细资料

threshold

protected float threshold

classes

protected List<String> classes

applyRatio

protected boolean applyRatio

removePadding

protected boolean removePadding

pipeline

protected ai.djl.translate.Pipeline pipeline

width

protected int width

height

protected int height

构造器详细资料

YoloV5FaceTranslator

protected YoloV5FaceTranslator(YoloV5FaceTranslator.Builder builder)

Constructs an ImageTranslator with the provided builder.

参数:

builder - the data to build with

方法详细资料

builder

public static YoloV5FaceTranslator.Builder builder()

Creates a builder to build a YoloV8Translator with specified arguments.

返回:

a new builder

builder

public static YoloV5FaceTranslator.Builder builder(Map<String,?> arguments)

Creates a builder to build a YoloV8Translator with specified arguments.

参数:

arguments - arguments to specify builder options

返回:

a new builder

processInput

public ai.djl.ndarray.NDList processInput(ai.djl.translate.TranslatorContext ctx,
                                          ai.djl.modality.cv.Image input)
                                   throws Exception

指定者:

processInput 在接口中 ai.djl.translate.PreProcessor<ai.djl.modality.cv.Image>

抛出:

Exception

processOutput

public ai.djl.modality.cv.output.DetectedObjects processOutput(ai.djl.translate.TranslatorContext ctx,
                                                               ai.djl.ndarray.NDList list)
                                                        throws Exception

指定者:

processOutput 在接口中 ai.djl.translate.PostProcessor<ai.djl.modality.cv.output.DetectedObjects>

抛出:

Exception

processFromBoxOutput

protected ai.djl.modality.cv.output.DetectedObjects processFromBoxOutput(int imageWidth,
                                                                         int imageHeight,
                                                                         ai.djl.ndarray.NDList list,
                                                                         float scale)

nms

protected ai.djl.modality.cv.output.DetectedObjects nms(int imageWidth,
                                                        int imageHeight,
                                                        List<ai.djl.modality.cv.output.Landmark> boxes,
                                                        List<Integer> classIds,
                                                        List<Float> scores,
                                                        float scale)