bounding box sort clockwise function is incorrect

[wybryan]

Describe the bug

from mmocr.utils.box_util import sort_points poly = [[1668,1748],[1704,1749],[1742,1898],[1705,1898]] sorted_pts = order_points(np.array(poly)) print(sorted_pts)

the MMOCR function sort_points claims to sort arbitory points in clockwise order.

BUT the above code produce sorted points as [1704, 1749] [1668, 1748] [1705, 1898] [1742, 1898]

which is completely incorrect. The expected order should be:

[1668,1748] [1704,1749] [1742,1898] [1705,1898]

Reproduction

1. What command or script did you run?

from mmocr.utils.box_util import sort_points poly = [[1668,1748],[1704,1749],[1742,1898],[1705,1898]] sorted_pts = order_points(np.array(poly)) print(sorted_pts)

The MMOCR version used is 0.6.0.

Environment

1. Please run python mmocr/utils/collect_env.py to collect necessary environment information and paste it here.

sys.platform: linux Python: 3.9.12 (main, Apr 5 2022, 06:56:58) [GCC 7.5.0] CUDA available: False GCC: gcc (Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0 PyTorch: 1.9.0+cu111 PyTorch compiling details: PyTorch built with:

• GCC 7.3
• C++ Version: 201402
• Intel(R) oneAPI Math Kernel Library Version 2021.4-Product Build 20210904 for Intel(R) 64 architecture applications
• Intel(R) MKL-DNN v2.1.2 (Git Hash 98be7e8afa711dc9b66c8ff3504129cb82013cdb)
• OpenMP 201511 (a.k.a. OpenMP 4.5)
• NNPACK is enabled
• CPU capability usage: AVX2
• Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, CUDA_VERSION=11.1, CUDNN_VERSION=8.0.5, CXX_COMPILER=/opt/rh/devtoolset-7/root/usr/bin/c++, CXX_FLAGS= -Wno-deprecated -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -fopenmp -DNDEBUG -DUSE_KINETO -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DSYMBOLICATE_MOBILE_DEBUG_HANDLE -O2 -fPIC -Wno-narrowing -Wall -Wextra -Werror=return-type -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-sign-compare -Wno-unused-parameter -Wno-unused-variable -Wno-unused-function -Wno-unused-result -Wno-unused-local-typedefs -Wno-strict-overflow -Wno-strict-aliasing -Wno-error=deprecated-declarations -Wno-stringop-overflow -Wno-psabi -Wno-error=pedantic -Wno-error=redundant-decls -Wno-error=old-style-cast -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Wno-stringop-overflow, LAPACK_INFO=mkl, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, PERF_WITH_AVX512=1, TORCH_VERSION=1.9.0, USE_CUDA=ON, USE_CUDNN=ON, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=OFF, USE_NCCL=ON, USE_NNPACK=ON, USE_OPENMP=ON,

TorchVision: 0.10.0+cu111 OpenCV: 4.5.4 MMCV: 1.5.3 MMCV Compiler: GCC 7.3 MMCV CUDA Compiler: 11.1 MMOCR: 0.6.0+

Bug fix

I have created a PR for it regarding the fix.

[wybryan]

@xinke-wang Hi,

I have done a PR fix at https://github.com/open-mmlab/mmocr/pull/1146,

let me know what you think, cheers.

[gaotongxiao]

Thanks! Since this PR works like a patch for a special case, I'd like to start from this discussion and invite more input from our community to see if we can finally achieve an optimum solution.

[gaotongxiao]

The ordering depends on the coordinate system that you are working on. We finally applied the fix at https://github.com/open-mmlab/mmocr/pull/1205, which is based on the Cartesian coordinate system. Simply reversing the points should address your issue. Let us know if there is any other problem.