zoo.models.image.objectdetection package

Submodules

zoo.models.image.objectdetection.object_detector module

class zoo.models.image.objectdetection.object_detector.DecodeOutput(bigdl_type='float')

Bases: zoo.feature.image.imagePreprocessing.ImagePreprocessing

Decode the detection output The output of the model prediction is a 1-dim tensor The first element of tensor is the number(K) of objects detected, followed by [label score x1 y1 x2 y2] X K For example, if there are 2 detected objects, then K = 2, the tensor may looks like `2, 1, 0.5, 10, 20, 50, 80, 3, 0.3, 20, 10, 40, 70` After decoding, it returns a 2-dim tensor, each row represents a detected object ` 1, 0.5, 10, 20, 50, 80 3, 0.3, 20, 10, 40, 70 `

class zoo.models.image.objectdetection.object_detector.ImInfo(bigdl_type='float')

Bases: zoo.feature.image.imagePreprocessing.ImagePreprocessing

Generate imInfo imInfo is a tensor that contains height, width, scaleInHeight, scaleInWidth

class zoo.models.image.objectdetection.object_detector.ObjectDetector(bigdl_type='float')

Bases: zoo.models.image.common.image_model.ImageModel

A pre-trained object detector model.

:param model_path The path containing the pre-trained model

static load_model(path, weight_path=None, bigdl_type='float')

Load an existing object detection model (with weights).

# Arguments path: The path to save the model. Local file system, HDFS and Amazon S3 are supported.

HDFS path should be like ‘hdfs://[host]:[port]/xxx’. Amazon S3 path should be like ‘s3a://bucket/xxx’.

class zoo.models.image.objectdetection.object_detector.ScaleDetection(bigdl_type='float')

Bases: zoo.feature.image.imagePreprocessing.ImagePreprocessing

If the detection is normalized, for example, ssd detected bounding box is in [0, 1], need to scale the bbox according to the original image size. Note that in this transformer, the tensor from model output will be decoded, just like DecodeOutput

class zoo.models.image.objectdetection.object_detector.Visualizer(label_map, thresh=0.3, encoding='png', bigdl_type='float')

Bases: zoo.feature.image.imagePreprocessing.ImagePreprocessing

Visualizer is a transformer to visualize the detection results (tensors that encodes label, score, boundingbox) You can call image_frame.get_image() to get the visualized results

zoo.models.image.objectdetection.object_detector.read_coco_label_map()

load coco label map

zoo.models.image.objectdetection.object_detector.read_pascal_label_map()

load pascal label map

Module contents