Secondary detector

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

wrong title, multiple detectors work in Deepstream ?

I have downloaded “facenet-120.tar.gz” “https://nvidia.box.com/shared/static/wjitc00ef8j6shjilffibm6r2xxcpigz.gz”

In single_source.txt

I added this at the bottom, this file also have [primary-gie]

[primary-gie]

enable=1
#model-engine-file=../../models/Secondary_CarColor/resnet18.caffemodel_b16_int8.engine
labelfile-path=../../models/facenet-120/class_labels.txt
batch-size=8
gpu-id=0
gie-unique-id=6
operate-on-gie-id=2
operate-on-class-ids=0;
config-file=face.txt

[property]
gpu-id=0
net-scale-factor=0.0039215697906911373
model-file=../../models/facenet-120/snapshot_iter_24000.caffemodel
proto-file=../../models/facenet-120/deploy.prototxt
#model-engine-file=../../models/Primary_Detector/resnet10.caffemodel_b30_int8.engine
labelfile-path=../../models/facenet-120/class_labels.txt
#int8-calib-file=../../models/Primary_Detector/cal_trt.bin
batch-size=16
process-mode=1
model-color-format=0
## 0=FP32, 1=INT8, 2=FP16 mode
#network-mode=2
num-detected-classes=1
interval=0
gie-unique-id=1
output-blob-names=coverage/sig;bbox/regressor
#parse-bbox-func-name=NvDsInferParseCustomResnet
#custom-lib-path=/path/to/libnvdsparsebbox.so
#enable-dbscan=1

[class-attrs-all]
threshold=0.1
group-threshold=1
## Set eps=0.7 and minBoxes for enable-dbscan=1
eps=0.2
#minBoxes=3
roi-top-offset=0
roi-bottom-offset=0
detected-min-w=0
detected-min-h=0
detected-max-w=0
detected-max-h=0

Opening in BLOCKING MODE 
Creating LL OSD context new
gstnvtracker: Loading low-level lib at /opt/nvidia/deepstream/deepstream-4.0/lib/libnvds_mot_klt.so
gstnvtracker: Optional NvMOT_RemoveStreams not implemented
gstnvtracker: Batch processing is OFF
0:00:03.873237017 17360     0x203e42a0 WARN                 nvinfer gstnvinfer.cpp:515:gst_nvinfer_logger:<primary_gie_classifier> NvDsInferContext[UID 6]:checkEngineParams(): Could not find output layer 'coverage/sig' in engine
0:00:03.873341502 17360     0x203e42a0 WARN                 nvinfer gstnvinfer.cpp:515:gst_nvinfer_logger:<primary_gie_classifier> NvDsInferContext[UID 6]:checkEngineParams(): Could not find output layer 'bbox/regressor' in engine
2

after reading forums I have added, still same issue,

process-mode=1

according to prototxt, below are the correct output layers

output-blob-names=coverage/sig;bbox/regressor
3

Hi,

FaceNet use the DetectNet architecture but trained on FDDB database.
Since the output definition is different from the resnet-10, you will need to update the bounding box parser as well.

Please check this GitHub for the source code of DetectNet bbox parser:
[url]https://github.com/AastaNV/DeepStream/tree/master/parser_detectnet[/url]

Thanks.