Modified Caffe parser to support the new dnn engine #26208
Now the Caffe parser supports both the old and the new engine. It can be selected using newEngine argument in PopulateNet.
All cpu Caffe tests work fine except:
- Test_Caffe_nets.Colorization
- Test_Caffe_layers.FasterRCNN_Proposal
Both these tests doesn't work because of the bug in the new net.forward function. The function takes the name of the desired target last layer, but uses this name as the name of the desired output tensor.
Also Colorization test contains a strange model with a Silence layer in the end, so it doesn't have outputs. The old parser just ignored it. I think, the proper solution is to run this model until the (number_of_layers - 2) layer using proper net.forward arguments in the test.
### Pull Request Readiness Checklist
See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request
- [x] I agree to contribute to the project under Apache 2 License.
- [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV
- [ ] The PR is proposed to the proper branch
- [ ] There is a reference to the original bug report and related work
- [x] There is accuracy test, performance test and test data in opencv_extra repository, if applicable
Patch to opencv_extra has the same branch name.
- [x] The feature is well documented and sample code can be built with the project CMake
Modified TFLite parser for the new dnn engine #26330
The new dnn graph is creating just by defining input and output names of each layer.
Some TFLite layers has fused activation, which doesn't have layer name and input and output names. Also some layers require additional preprocessing layers (e.g. NHWC -> NCHW). All these layers should be added to the graph with some unique layer and input and output names.
I solve this problem by adding additionalPreLayer and additionalPostLayer layers.
If a layer has a fused activation, I add additionalPostLayer and change input and output names this way:
**original**: conv_relu(conv123, conv123_input, conv123_output)
**new**: conv(conv123, conv123_input, conv123_output_additional_post_layer) + relu(conv123_relu, conv1_output_additional_post_layer, conv123_output)
If a layer has additional preprocessing layer, I change input and output names this way:
**original**: permute_reshape(reshape345, reshape345_input, reshape345_output)
**new**: permute(reshape345_permute, reshape345_input, reshape345_input_additional_pre_layer) + reshape(reshape345, reshape345_input_additional_pre_layer, reshape345_output)
### Pull Request Readiness Checklist
See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request
- [x] I agree to contribute to the project under Apache 2 License.
- [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV
- [x] The PR is proposed to the proper branch
- [ ] There is a reference to the original bug report and related work
- [x] There is accuracy test, performance test and test data in opencv_extra repository, if applicable
Patch to opencv_extra has the same branch name.
- [x] The feature is well documented and sample code can be built with the project CMake
New dnn engine #26056
This is the 1st PR with the new engine; CI is green and PR is ready to be merged, I think.
Merge together with https://github.com/opencv/opencv_contrib/pull/3794
---
**Known limitations:**
* [solved] OpenVINO is temporarily disabled, but is probably easy to restore (it's not a deal breaker to merge this PR, I guess)
* The new engine does not support any backends nor any targets except for the default CPU implementation. But it's possible to choose the old engine when loading a model, then all the functionality is available.
* [Caffe patch is here: #26208] The new engine only supports ONNX. When a model is constructed manually or is loaded from a file of different format (.tf, .tflite, .caffe, .darknet), the old engine is used.
* Even in the case of ONNX some layers are not supported by the new engine, such as all quantized layers (including DequantizeLinear, QuantizeLinear, QLinearConv etc.), LSTM, GRU, .... It's planned, of course, to have full support for ONNX by OpenCV 5.0 gold release. When a loaded model contains unsupported layers, we switch to the old engine automatically (at ONNX parsing time, not at `forward()` time).
* Some layers , e.g. Expat, are only partially supported by the new engine. In the case of unsupported flavours it switches to the old engine automatically (at ONNX parsing time, not at `forward()` time).
* 'Concat' graph optimization is disabled. The optimization eliminates Concat layer and instead makes the layers that generate tensors to be concatenated to write the outputs to the final destination. Of course, it's only possible when `axis=0` or `axis=N=1`. The optimization is not compatible with dynamic shapes since we need to know in advance where to store the tensors. Because some of the layer implementations have been modified to become more compatible with the new engine, the feature appears to be broken even when the old engine is used.
* Some `dnn::Net` API is not available with the new engine. Also, shape inference may return false if some of the output or intermediate tensors' shapes cannot be inferred without running the model. Probably this can be fixed by a dummy run of the model with zero inputs.
* Some overloads of `dnn::Net::getFLOPs()` and `dnn::Net::getMemoryConsumption()` are not exposed any longer in wrapper generators; but the most useful overloads are exposed (and checked by Java tests).
* [in progress] A few Einsum tests related to empty shapes have been disabled due to crashes in the tests and in Einsum implementations. The code and the tests need to be repaired.
* OpenCL implementation of Deconvolution is disabled. It's very bad and very slow anyway; need to be completely revised.
* Deconvolution3D test is now skipped, because it was only supported by CUDA and OpenVINO backends, both of which are not supported by the new engine.
* Some tests, such as FastNeuralStyle, checked that the in the case of CUDA backend there is no fallback to CPU. Currently all layers in the new engine are processed on CPU, so there are many fallbacks. The checks, therefore, have been temporarily disabled.
---
- [x] I agree to contribute to the project under Apache 2 License.
- [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV
- [x] The PR is proposed to the proper branch
- [ ] There is a reference to the original bug report and related work
- [ ] There is accuracy test, performance test and test data in opencv_extra repository, if applicable
Patch to opencv_extra has the same branch name.
- [ ] The feature is well documented and sample code can be built with the project CMake
Objc binding
* Initial work on Objective-C wrapper
* Objective-C generator script; update manually generated wrappers
* Add Mat tests
* Core Tests
* Imgproc wrapper generation and tests
* Fixes for Imgcodecs wrapper
* Miscellaneous fixes. Swift build support
* Objective-C wrapper build/install
* Add Swift wrappers for videoio/objdetect/feature2d
* Framework build;iOS support
* Fix toArray functions;Use enum types whenever possible
* Use enum types where possible;prepare test build
* Update test
* Add test runner scripts for iOS and macOS
* Add test scripts and samples
* Build fixes
* Fix build (cmake 3.17.x compatibility)
* Fix warnings
* Fix enum name conflicting handling
* Add support for document generation with Jazzy
* Swift/Native fast accessor functions
* Add Objective-C wrapper for calib3d, dnn, ml, photo and video modules
* Remove IntOut/FloatOut/DoubleOut classes
* Fix iOS default test platform value
* Fix samples
* Revert default framework name to opencv2
* Add converter util functions
* Fix failing test
* Fix whitespace
* Add handling for deprecated methods;fix warnings;define __OPENCV_BUILD
* Suppress cmake warnings
* Reduce severity of "jazzy not found" log message
* Fix incorrect #include of compatibility header in ios.h
* Use explicit returns in subscript/get implementation
* Reduce minimum required cmake version to 3.15 for Objective-C/Swift binding
