How can I get the confidence of model license-plate-recognition-barrier-0001?

[zhiyong]

I build a DLStreamer license plate recognition demo, the models are from security_barrier_camera_demo, after making the pipeline running,  some results appear.

but I can't get the confidence from the license-plate-recognition-barrier-0001 model(always 0),  the results are as follow:

Object bounding box 561,544,50,12,
Attribute license_plate
label= <Jiangsu>EF83Y2
confidence= 0
Object bounding box 486,396,220,177,
Attribute color
label= white
confidence= 0.999989
Attribute type
label= bus
confidence= 0.929861

so how can I get the correct confidence for plate recognition?

ps:

void PrintMeta(GstBuffer *buffer, GstCaps *caps) { // simple function to display some tensor information

GVA::VideoFrame video_frame(buffer, caps);

std::vector<GVA::RegionOfInterest> regions = video_frame.regions();

for (GVA::RegionOfInterest &roi : regions) { // iterate by regions of interest attached to this video frame

auto bbox = roi.rect(); // get bounding box information

std::cout << "Object bounding box " << bbox.x << "," << bbox.y << "," << bbox.w << "," << bbox.h << ","

<< std::endl;

for (const GVA::Tensor &tensor : roi.tensors()) { // iterate by tensors attached to this region of interest

if (tensor.is_detection())

continue; // detection tensor doesn't contain classification result and hence doesn't contain label

// print some tensor information

std::cout << " Attribute " << tensor.name() << std::endl;

std::cout << " label= " << tensor.label() << std::endl;

std::cout << " confidence= " << tensor.confidence() << std::endl;

}

}

}

[Vladimir_Dudnik]

@zhiyong  according to license-plate-recognition-barrier-0001 description, model output does not contain confidence.

Encoded vector of floats, name: decode, shape: 1, 88, 1, 1. Each float is an integer number encoding a character according to this dictionary:

    0 0
    1 1
    2 2
    3 3
    4 4
    5 5
    6 6
    7 7
    8 8
    9 9
    10 <Anhui>
    11 <Beijing>
    12 <Chongqing>
    13 <Fujian>
    14 <Gansu>
    15 <Guangdong>
    16 <Guangxi>
    17 <Guizhou>
    18 <Hainan>
    19 <Hebei>
    20 <Heilongjiang>
    21 <Henan>
    22 <HongKong>
    23 <Hubei>
    24 <Hunan>
    25 <InnerMongolia>
    26 <Jiangsu>
    27 <Jiangxi>
    28 <Jilin>
    29 <Liaoning>
    30 <Macau>
    31 <Ningxia>
    32 <Qinghai>
    33 <Shaanxi>
    34 <Shandong>
    35 <Shanghai>
    36 <Shanxi>
    37 <Sichuan>
    38 <Tianjin>
    39 <Tibet>
    40 <Xinjiang>
    41 <Yunnan>
    42 <Zhejiang>
    43 <police>
    44 A
    45 B
    46 C
    47 D
    48 E
    49 F
    50 G
    51 H
    52 I
    53 J
    54 K
    55 L
    56 M
    57 N
    58 O
    59 P
    60 Q
    61 R
    62 S
    63 T
    64 U
    65 V
    66 W
    67 X
    68 Y
    69 Z

[Wan_Intel]

Hi Vladimir_Dudnik,

Thank you for sharing in the OpenVINO™ community!

Hi Zhiyong,

Thank you for your question.

If you need any additional information from Intel, please submit a new question as this thread is no longer being monitored.

Regards,

Wan

[MaximD]

Hello! Could you share the pipeline used in your approach? I mean the argument after gst-lauch command.