[FFmpeg-devel] [PATCH 6/6] lavfi/dnn_classify: add filter dnn_classify for classification based on detection bounding boxes

[Andreas Rheinhardt]

Guo, Yejun:
> 
> 
>> -----Original Message-----
>> From: ffmpeg-devel <ffmpeg-devel-bounces at ffmpeg.org> On Behalf Of
>> Andreas Rheinhardt
>> Sent: 2021年4月26日 9:34
>> To: ffmpeg-devel at ffmpeg.org
>> Subject: Re: [FFmpeg-devel] [PATCH 6/6] lavfi/dnn_classify: add filter
>> dnn_classify for classification based on detection bounding boxes
>>
>> Guo, Yejun:
>>>
>>>
>>>> -----Original Message-----
>>>> From: Guo, Yejun <yejun.guo at intel.com>
>>>> Sent: 2021年4月18日 18:08
>>>> To: ffmpeg-devel at ffmpeg.org
>>>> Cc: Guo, Yejun <yejun.guo at intel.com>
>>>> Subject: [PATCH 6/6] lavfi/dnn_classify: add filter dnn_classify for
>>>> classification based on detection bounding boxes
>>>>
>>>> classification is done on every detection bounding box in frame's side
>> data,
>>>> which are the results of object detection (filter dnn_detect).
>>>>
>>>> Please refer to commit log of dnn_detect for the material for detection,
>>>> and see below for classification.
>>>>
>>>> - download material for classifcation:
>>>> wget
>>>>
>> https://github.com/guoyejun/ffmpeg_dnn/raw/main/models/openvino/20
>>>> 21.1/emotions-recognition-retail-0003.bin
>>>> wget
>>>>
>> https://github.com/guoyejun/ffmpeg_dnn/raw/main/models/openvino/20
>>>> 21.1/emotions-recognition-retail-0003.xml
>>>> wget
>>>>
>> https://github.com/guoyejun/ffmpeg_dnn/raw/main/models/openvino/20
>>>> 21.1/emotions-recognition-retail-0003.label
>>>>
>>>> - run command as:
>>>> ./ffmpeg -i cici.jpg -vf
>>>>
>> dnn_detect=dnn_backend=openvino:model=face-detection-adas-0001.xml:
>>>>
>> input=data:output=detection_out:confidence=0.6:labels=face-detection-ad
>>>>
>> as-0001.label,dnn_classify=dnn_backend=openvino:model=emotions-recog
>>>>
>> nition-retail-0003.xml:input=data:output=prob_emotion:confidence=0.3:la
>>>> bels=emotions-recognition-retail-0003.label:target=face,showinfo -f null
>> -
>>>>
>>>> We'll see the detect&classify result as below:
>>>> [Parsed_showinfo_2 @ 0x55b7d25e77c0]   side data - detection
>> bounding
>>>> boxes:
>>>> [Parsed_showinfo_2 @ 0x55b7d25e77c0] source:
>>>> face-detection-adas-0001.xml, emotions-recognition-retail-0003.xml
>>>> [Parsed_showinfo_2 @ 0x55b7d25e77c0] index: 0,  region: (1005, 813)
>> ->
>>>> (1086, 905), label: face, confidence: 10000/10000.
>>>> [Parsed_showinfo_2 @ 0x55b7d25e77c0]            classify:  label:
>>>> happy, confidence: 6757/10000.
>>>> [Parsed_showinfo_2 @ 0x55b7d25e77c0] index: 1,  region: (888, 839)
>> ->
>>>> (967, 926), label: face, confidence: 6917/10000.
>>>> [Parsed_showinfo_2 @ 0x55b7d25e77c0]            classify:  label:
>>>> anger, confidence: 4320/10000.
>>>>
>>>> Signed-off-by: Guo, Yejun <yejun.guo at intel.com>
>>>> ---
>>>>  configure                     |   1 +
>>>>  doc/filters.texi              |  36 ++++
>>>>  libavfilter/Makefile          |   1 +
>>>>  libavfilter/allfilters.c      |   1 +
>>>>  libavfilter/vf_dnn_classify.c | 330
>>>> ++++++++++++++++++++++++++++++++++
>>>>  5 files changed, 369 insertions(+)
>>>>  create mode 100644 libavfilter/vf_dnn_classify.c
>>>>
>>>> +
>>>> +AVFilter ff_vf_dnn_classify = {
>>>> +    .name          = "dnn_classify",
>>>> +    .description   = NULL_IF_CONFIG_SMALL("Apply DNN classify
>> filter
>>>> to the input."),
>>>> +    .priv_size     = sizeof(DnnClassifyContext),
>>>> +    .init          = dnn_classify_init,
>>>> +    .uninit        = dnn_classify_uninit,
>>>> +    .query_formats = dnn_classify_query_formats,
>>>> +    .inputs        = dnn_classify_inputs,
>>>> +    .outputs       = dnn_classify_outputs,
>>>> +    .priv_class    = &dnn_classify_class,
>>>> +    .activate      = dnn_classify_activate,
>>>> +};
>>>
>>> I've locally added 'const' for AVFilter ff_vf_dnn_classify, any other
>> comment? thanks.
>>>
>> If you did this, then this filter may only be added after the bump.
> 
> thanks, I'll remove/add 'const' by checking the bump when it can be pushed.
> 
> btw, any rough estimate when the bump will happen? thanks.
> 
Really soon if no issues are detected any more. Probably done in two days.

- Andreas