I am trying to use the code you provided, but I have some questions...

[Bahramudin]

First of all thanks for this effort!!

Secondly, I have am trying to train my own image dataset with Tensorflow, I have already done according to the README file (prepared images for 2 objects each 200, and run the train.sh commands and everything go well), it has created a folder calls tf_file and one inception.

The question is that, then how can I use the trained data in OpenCV for detecting the objects?
for example, the OpenCV DNN module support Tensorflow, when I use the ssd_mobilenet_v1_coco_11_06_2017 trained data OpenCV needs two files:

1. frozen_inference_graph.pb
2. ssd_mobilenet_v1_coco.pbtxt

which can be something like this:

String basePath = "D:\\ssd_mobilenet_v1_coco_11_06_2017\\";
String model = basePath + "frozen_inference_graph.pb";
String config = basePath + "ssd_mobilenet_v1_coco.pbtxt";
Net net = Dnn.readNetFromTensorflow(model, config);

And then I am able to detect the objects.

Now how can I get this result by training my own dataset?

When I use the .pb file in OpenCV it says:

OpenCV Error: Unspecified error (Unknown layer type DecodeJpeg in op DecodeJpeg) in cv::dnn::experimental_dnn_v3::anonymous-namespace'::TFImporter::populateNet, file C:\opencv\modules\dnn\src\tensorflow\tf_importer.cpp, line 1487 Exception in thread "main" CvException [org.opencv.core.CvException: cv::Exception: E:\git_folder\opencv\modules\dnn\src\tensorflow\tf_importer.cpp:1487: error: (-2) Unknown layer type DecodeJpeg in op DecodeJpeg in function cv::dnn::experimental_dnn_v3::anonymous-namespace'::TFImporter::populateNet
]
at org.opencv.dnn.Dnn.readNetFromTensorflow_1(Native Method)
at org.opencv.dnn.Dnn.readNetFromTensorflow(Dnn.java:162)

Thanks in Advance!!