According to experiment, for sorting std::vector, when the target to be sorted is less than 12.5% of the whole vector, apply partial_sort instead of sort benefits at most x8 speedup.

Typically within detection out layer, the output is far less than the proposals after NMS; for example, for faster RCNN model coming from TensorFlow, there are 90 (# of classes) * 100 proposals but outputs 100 outcomes only. Sometimes even after NMS, the proposals is still far largen than the number to keep. Ex, for maskRCNN, we keep top 100 but proposal after NMS is 1292, so the ratio is 100/1292 ~ 8% and partial sort is faster x 2 times than sort. In the case if we keep top 10 only, the ratio is 10/1292 < 1% so partial sort benefits more.
As the result, in this case, apply partial sort makes it faster x8 times than applying std::sort().

Architecture:        x86_64
CPU op-mode(s):      32-bit, 64-bit
Byte Order:          Little Endian
CPU(s):              12
On-line CPU(s) list: 0-11
Thread(s) per core:  2
Core(s) per socket:  6
Socket(s):           1
NUMA node(s):        1
Vendor ID:           GenuineIntel
CPU family:          6
Model:               158
Model name:          Intel(R) Core(TM) i7-9850H CPU @ 2.60GHz
Stepping:            13
CPU MHz:             3699.935
CPU max MHz:         4600.0000
CPU min MHz:         800.0000
BogoMIPS:            5199.98
Virtualization:      VT-x
L1d cache:           32K
L1i cache:           32K
L2 cache:            256K
L3 cache:            12288K
NUMA node0 CPU(s):   0-11

Following table shows the speedup of the partial sort ratio; from it we can see when ratio < 25%, partial sort is faster than sort.
To make it safe, 12.5% is chosen.

Pull Request Readiness Checklist

See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request

• I agree to contribute to the project under Apache 2 License.
• To the best of my knowledge, the proposed patch is not based on a code under GPL or other license that is incompatible with OpenCV
• The PR is proposed to proper branch
• There is reference to 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