# Image segmentation and encode with TFLite

Source: [https://docs.qualcomm.com/doc/80-70014-50/topic/single-camera-stream-with-image-segmentation-and-encode.html](https://docs.qualcomm.com/doc/80-70014-50/topic/single-camera-stream-with-image-segmentation-and-encode.html)

The use case uses the `deeplabv3_resnet50` TFLite model to compose the
        semantic segmentations and original video stream, encode this stream, and then multiplex it
        in a MP4 container.

Use the following command to execute the use
            case:

    gst-launch-1.0 -e --gst-debug=2 \
    qtiqmmfsrc name=camsrc ! video/x-raw\(memory:GBM\),format=NV12,width=1280,height=720,framerate=30/1,compression=ubwc ! queue ! tee name=split \
    split. ! queue ! qtivcomposer name=mixer sink_1::dimensions="<1920,1080>" sink_1::alpha=0.5 ! queue ! video/x-raw\(memory:GBM\),format=NV12,width=1920,height=1080,interlace-mode=progressive,colorimetry=bt601 ! v4l2h264enc capture-io-mode=5 output-io-mode=5 ! h264parse ! queue ! mp4mux ! queue ! filesink location=/opt/video.mp4 \
    split. ! queue ! qtimlvconverter ! queue ! qtimltflite delegate=external external-delegate-path=libQnnTFLiteDelegate.so external-delegate-options="QNNExternalDelegate,backend_type=htp;" model=/opt/deeplabv3_resnet50.tflite ! queue ! qtimlvsegmentation module=deeplab-argmax labels=/opt/deeplabv3_resnet50.labels ! video/x-raw,width=256,height=144 ! queue ! mixer.Copy to clipboard

To stop the use case, press CTRL + C.

Figure : Pipeline for segmentation and encode with qtivcomposer
            
            ![](data:image/png;base64,UklGRn4pAABXRUJQVlA4THIpAAAvzkNbAGpR3LaNI+0/dpLr5RsRE0Cn3NUYdIp5Yl6xi7Q74Q91s87VonyjJRgBG5uMp2mUN/ViYPdiqFzebv3/rtxGDBHKSyFDhAwRMkSIUF4TMkSIcEKECBEiRMgQIcIJETJEiBAhMx4NMXfee/e+++bxL+CeI9uNcmSMXgeyoXK08iJmakK5GrARSlANFj0wlqMZVqAO1IiacBMykkMjcrQFyK6557ACNbB1uNTLDXjObeRhXbRleEMz5wCvkjUvXW/QwXrv5dDI2khnY7/y3ks1rJtG1g2NXLS+C0WTvXVNMGO03r1CGMmhgdsDDGzb6mAffvifz4cwDIdhGIZhGA7DYRgOwzAchh+GYTiczQW2bXl18GB4/k84DIfhMBwOh2E4DIfDYTgMw/DDMPwwzN7+Q2IkyVEUhdvFaKYZOLsr5qs52f7NkTTLvWPC98fyLGc5yzFhlmd5lrM8y1m+y16+y3c5y1nOcnYD0j8is7uqvybPZxCQNv0MCRDjVlmRtIvvavxI/hYmCP7WJRQC8YOwo+1pEgQi4HGn7WgQiICAyJ8t6UwQEKiJIn0rEAQSUhRlXCFIyEDQziUUM3MCxPHq5wNLbdvy6A7g/38YDofDYbhDGA7DHcIwDMPwxeGHYRiGYZaQIPn/1kiz/5IgSZLbpiAhBBq0a3pmYV3U8R11A5rD/Mf8x/y3iViRguoNwyN0kJodGqg7QHDvy4U5G+kkj0LKV622QANhjaGAEgT5KQYME9BApQfBvb8zk3AjZKTTbBSWG+i7amEFDZRblDwgxEB+1n5AqYQGqgJSnNi4NybcCBnpNBuFFRb6rlpaG+v/gPRohdjemxVStyGDnqumnfV/QXY0Q2rvzQtPpyGEjqvG/Mf8x/zXUgDVb3oMIfVE6PkHzohVzzIwzf7+GmXWb3aU9UGV1bgfs6qh4nnJ43KWwYPp+e2M6vdc9EH/9bq1rX6To/+6XaD/fh2Ce2VYsWA1VDxWLUTsvx5XfTCtTB+P57cxCuj3WVShyrs/PeZWv7lR5H96/p/ZQL7KwZanN6taWvzJ9iqHmlb2enWzA1jZ1tObaQs1dOrVf2Hj64F63rBQHwGPwZk+qoMjj+frjCoz+5Tnk8f3QLxffouZkf918LfJznpeDz57xLw2Z03+eYZQzytLLJ9ajQvCH6zgxsp/E/9dlopMf/8LAvora27tf5fTzPKwcmD2N0EQBOsaLH92vQ+C4Hf6HEMduyajD3p2tp7cXxvMzupQ3hDwiDB73hzx+gw1hkcez9cXVXwB+xw+5fnk8T0Pn2Fm+Yv3v/Yyqde/M3IYM33GJKuyz6ibQ++zEM4xhMpnyfqtCyKpWZCZBWRJfxOZuD/Z74NZ/uf3fbRyBszKMqM/kGy3pgKSdH8bpcmBFgxlmQc0H+C/upXNnDxxvx4oQ5wTqLZiCjCctc4JQ5zg3ivGbB3Ewcfzq6ZhcmU8JmgF/6p4H7GwUbW0qkbdYlv9PkX9qFX72Xi2uMzhERHES7qJEcewDqz3EGm9k+wz4dqKQJyAxQn0Hk6sGGYKEcNMWytO/GXvIa5wXdErsbOr82XHpfdmQQwHZsmCoDOJD9Q46ypaDvivYuectBhHQS4iX2Baub+UU+rELvoEbPOeDLzr6R7OmgvbL0wSKEuelm/n+Lf9iqwp304YmDbZqnybCadlymA7nZk+WZNMypWMq7KcMFi/Yqat9SCHcFUR9wlaUPCsi5bPkGRWNhlw38rONNDvUdQRrYSPokWfZ6Mpg6uymVezfH78u6RI8+qpzpoEinyW1j6TLElqTHjLs0nelBXLppwlwK7JsreGGTBvylkirbFKmjmDLVdlWTFQP33sluuv7xH4TSvkaogYq9Jpx2HGA7bTJgPq/fHo3VXrHCwZLWRJ6bQs58D1YcZgt6TM/pXTtwJBzcPb+oI2wt5VZISZNGXFYOC5KdNcWOKnLDch9R2tgWUQhR0txjKjsO8pml12V/Q7jWXTBWFHv4fx44HCvp8kR+o7WtQ4XxFPUfNeO5gNBSiszoptn+Cfpd74fRD0paRd0EfBmoFZFIRRsGSoah8zESeOZQ2V04E2EupL6vuAKCujoFvs+iCf0RqQDX2S5Ej9FX1h1JL1Q4FFLeMg2vcRLZdRT2HiUuq6vqMXtrymKAy6CZBut2FEf6mwTI/U93TcSRJtuvmvfnOL4LKJczCedBaAsIYuujKdtv8xJN7TSoB3VIkcguz91nOzdbhnl9KdW9Kx77f0FhDhe+BE02T7e1wd+9NcwDKhSTPEplEQRj+uWVIKw76j36kZvKaop3AlXOKHPAxS1C+UZrRPpC4ojVPqUjsP6CXHgjLZRzeuXvz458/xkk5Ty78n9JZ5SQdpz1VFiPMwZ5UnZJe1PuRKD9K52OqqiL7Csat+QUZd7W62+8zmC2ri52C0kuQdpV+BcEjQMLfONgtZ2yyOk7cuLhK6Mh1/xrMmqlhTCX6KoixeUhbnYXdddz0Dt7TBe2qYD7Ss+8ES4CUVqDu64+uQRpm8pfKxoqBkLmj2WNLi2mZ9tMu7PnHXfyDBFMcgZW7oZK/71tdzEXuFCC2l1NtS7xSlbN2UJRy2cPDojOvL51dCcXFHZdoXJ32X9YxKizwascypkaMXZmXJ90H2eKAutdOuS8CskBmd7Cbo+9vlVVBZ5P0nVw6wWbDPJF9Q48ww39E8LulTjgn9lVcL7JSloqWD5IvZZpHF8nFObwesndR9l4hraCJ9NK8d8rFb0tzJhl4sJC+WjPMQ+sDanEv0rDzFymQdZ6WHh5UtC7bc1qRqilGXy5qyjxLXfcgFTT9KnEcjhlp7rS2oRJv7zEKKpXEcJ8gN/pBj/5RnoowkOgKwp2DIsxuQuDVtZHwVIqPfBJaLMqW/+Haw+CFIOBpBUFD1KBktpKJCIHl3zMMuF7iUltOgz9gh5w0dYuceC8ry7h+fdm+TsaNwbU4levqLkpQoM6BF2pq8QqyQnUtVRY1zNc+dUePqcMStD8bzkZaQAzUObk2ZsELSIFrFFfWZuF0f5fIhmMafB8iaVh8l5j5kI0yX0gRGmNJ82gku8Z/cUSXMEAiq4nQM6VZSOgjyvk8gb2nilkT9h1ktQ0Qarye/ioNCpsq8G0EluqCJMY9ofxxsFDyPgpFpkFguVS6cYyYWvpGbEF63+l79iXiQFa2FwSZzk4wO8YRK11AjzJDHJT0LsyxpxdERGDARmdPCVrS2LXAMN8GJocij91wQhb9V7A7UH4/HHzqa1d3mPUKSMAqnTbg45/TXHmn0gu9T8WPmPgQ7WUfJRzltk0cTyGpEa8AeKBU2libDNhRtRCZ0K2jdgbJJUMUf31L2kY80Oh6PJwrGE1rbVlKqdkaYLMIc/8kbFEPuIbI7UlgcfqJiiPVFsrvwbzMKs3hJG5GSJqZn+5Wfku9bUUc0+oG0sg+K9WArZ6swGhrclU0OaZ5uYvWNHdHr6JHuw815EB71vA6mbjlEW4kPXm4GtGcB/bfjAybBFzCEu6OVzbonOiZLOt2t1+tlucrP4kaMBn8MK+Gjv7amuA/EsqEm317Kiq4+L5d9UKSPQyCTjpYAD3iyJuA19XMBVttTC5aGypAOy/KHYNEko6AYxlQ2XNEaBlO6HGIFvSQsNJQ6BmfJmhqWjxm9nAk/5zJODlR8NHiiUgDOMtav+i/6SfSShvAu2H10TnZJfdzmpgGrZ5qM6GVc9BKandKLtIBJZgApqepPkIqWjoEsKalyDClodx6kBff7VbdnsGTBS56xy3cFlSk1H517zKecvDbUDy51tqYmBf++ouDwhyeqHrPTcb8Po6BfxzNKRdKAGgH74viSjjtpGRx2OdgdaFIc9+G+C/r3+eKX64/OSbKDD+Noz2iRzMeCBDfBaOxcRcsX2sUOy7N384RGHH9ceUDSXe3c45zeAUrmv688KN5H9ceUbtm5OZ1cQwc4V9EHqHn79xXFOIymTubdNnssafNxPARJ1wepIIn6xLmMFitjCay2YX0OLllcSgdZUCoOBc3nVEgcT6gcR2EibhwGeX42OKMESxlchTVgrRX8sJ2KzCiNs200iZ2I+HKgnzgWEdgllc4lfZ+LtTYu6dnJjNZwbkMn6yWNF5Qaye2SXGL2oK1dUtgsgz5PKWyaY2Ds1gM688I9cQWdZmUfbOIlTYSlob4su20m/PqyoNsNF3ScLaMoE76k0+wQXN0IS/OGpfqxb9ZREGUupVGahNsB+EBXOVhSuirLKLqRJYWzMtrOpQ72A16oEtnQJ1sFwag5BGEiq57WsxMVzEf6kC6jbirpj91ytqdSkugfRLO7gA621bZtwSO6EddQ6k4Upk+z2WyCO5oNuKNsRXRIZ4MtkXxEt034YyXtYJMlbQS88Kwnl1IxYEazeNXTuhnR22GduD0okI66/mVnMdt3fdGEB5n8tVMC+fv3OST9dS7jz/uuP82tNL9mwkB16vt3xjOxsmP3Dtzsu36xsgwu913/MrX8PW73oC2qUbdP3weZ5L/1/WpxTMCy+XUJVdjJD33/KQMwO3b9ZSaoRy8Dlr/ORbLwzlZ0eun7l52wnRZ9t3/LQHIXdv3LCsD81F8dK9j8D/SVJPI9vP31RlhVFfyyn0KqXzf8069hf3V1FV2i+bUCS/PrdN4bKl2Uic3XfXfc2FZVWWbGsuW3xnoSbH5dDqh+rcQMMwWkh9+DwoIZ4DcAM8Cez8v1OY9hCNnW9Ll/Yw0N1tf6kVzM99449B4AtwDr97rPFVcF2LjZAj4PRQBDdeBkzcG+h3ekPY9N8Gz2n2gtLQOsDDADuFegNRzvPdZmNZLgLblUwZ7Z90tE+QwD5TM3oGVYHdqqZ9nC+i1vbwEwk1OPBA/avgGUVfVnQBWqajT3zZthh6EeIfb6KJuer/XjrH+GqjeOB5+6v/c99ONngIewAvoGaqxMY4V5Jtl/1vBm4GUOPTDEAjqcelbgZ8NAPbKGyhtWHa4UgjfM9Evp4cEnkWHt5qaGiukG08VQMyPx9Va/Ze4TvbJHjwwPanTlDAFT58Gr4OPj7/k686C+KfjV/X2NM/5bHVEmbFwEvAN1vsk/c+hZUips61UZqvlNo1fFlCV8O2v1tfDGouZmLkp/l/M95NI3ZB9R3+T+/vWLGik36wSvYOoU2W0FvYD8koPKIrh/FdzDWb6I5EGZ9ZUMnQ59Hy4iedBXNnKqF85frv9/+8//9p/mggxr0gremNxJK3ge5qqljWlKKXJM7pRSpMRctbQxTfiIJ039hI+SLt0nfCT2In7CR7fp9eTngtdrMcktbXtfzEKZLpo2Mr01H67BPNLLHV+At9XzT1eN+NdpSKa1RcRctWt8Q0ohl1X58M18nlXbIt2rVurXEmP8BCRdB5p5F6+Ej/vobqf7oMHsXFB7ZEhWRzQtsPoh2WzLCYJH4Fik5exK+Fjt4ZMfBPe19QzHdro462SzziNgrI4o90hH2bpdnKsfFZDvtunX5sHyA9ltJzH/Mf+xOE2BZiugWdnqQJltbzrlLTfZNhibt1tH19OKhN4tHDwYm9ZXACLIW8EyOF69Vc5xrNk+HJvQtRI+gmykESTfrXzbreg5JJtQyFT+Xlj+e/fYkFl8cDTywqgLWNMwuI6Kj7XViyHKBT1zseYpT8Hb6p75AOXGzwkdJ5UBKOfo+/MepSGKnj+NGUyg+acNZP5j/mP+a/mFqr+Dqp+BnxmkUma/Swfz0MLPANUyMKVsHVhNBYiIgNXPILZlvRJMobq9AatxuRi/fXlZJpYZ9ax4KROwDnXmtze4KEUVBaXCqg+KZETH098mfQaMT7Q/UTi1rIy8N830YXg0lH0w3qD3GJhv36tXQ/WijzVVAqiCT5RamUd7liUdICmdAAVuKagsK1QBZVVlBWBcVqDDI1PjwRR6w6oXfjhmAHlwEkA+BHndhfwo0jT8hmWy7YcwA8yw7f0bC+YWYPW4oOUW0HtYwwz3esFHc+j6YgyeXkNb+QJovKI15ptp7Jht3r2sqbKy+vVw6K8+ZQAmp74LD2OLMmwuo3Rc/LHrbxMAz6fu6jQBLvrYHmcvtGBYKEvS9ZxScdz+EpQAUHT53ZCMqC+XQbRzFXVleqJPbG8p+od0856KtKA/1pIGfdn8kRrb6sUeC3Y4RglYWcYnatySonWye0cNJKX08dZknzg3ocIeaRrH8i5IZE1LiTN6kdg1RT3uwnEcj/dRggs+Jq6Vl2AAS36iF5YlLcVKftUh7z5ZWVMlA5aOwX2I6XScrJIPNDWE7CrY3m4SiSWlIplOkzXNhPWC726tfaGdbWUVUsGwDTXCwKdg+ls0tbilCgMOMmDUg9/20TYIApMWaUgUvJvLkoY2wfbH8qIQ2XR0aMF2TrdDFttJR1ddFwXBHh72IS/oQ5XkL7QyYMScLX+ibdbQ4Xq32+X5eHDzRSAu21Jj/3W5KvLuj7VYDvt8U6VpdaL1RDJaxpDr7ftVcBSR+NaDPH/IYnFLKle0cCIufZniQpCEQ2qcAIAtac18R5+tiEi8pk08IJy04xeqrqN+KnkZ0dRkTsdExpc0se+orO0kiPKLPQpKpZWf6Dol6sMw7MONcEFdRwWDmVv7QqkM6KI+ogOkoSDs+o5WUlAqwJKisAvuWK5P1IUUZuCLOx4U82YKxaapb5qmNLYp0G7KcgLcq+rQbIcBd1lZzgHg+e3nZpc1uUcZz83ybQYweFKWaW5ZHy7wYHG4N+4H59lwrxDnBKpeM5aMDrFz0qq2hpM4mMoKESdDSTZshy9f0IHaKYJxPzY3rwez+pkNuRPDxXQydJnV68G+X2zhQvCLGo1+64iNkp0axUK1qUlzY5pmyvrZghp8+z0Qn/7EJKFRq4FUr3BFLyDzpORSeoDM8yQ4766f+Y/5r2U1OON5ADlpBfEynJESa6MFTEkJrOFc4AxMaRioJ9R34Mz2JdKWh7AUb747ZhrNM/JuxUp/38ccaGS3vbNlbfNgWcPkv57XzFnVHoBZR3W+WtI2rUjo4SU2cza8XMCss6ZEaPYINPsiNNsalnUP4eMjwKwr0jI0WwbLDu+6F5bVERUydTEw363r6YCsYfJ51Tl8v+/hffaxoZjd2R44K+GjAERHw2xgvHCeFaz1IoNvnOuDMbvjG441MPLqgHyTeVZw8fB8+YDMLhK6feNzUH6gkwaB2MdPkBw28gRVslidxaT3WaWXUzeLDUfjxeHgRt0tUfIa3kOR8LPUSX+6h+gctVSV9326h7LhH6OR933wNVjz++VYG738NS83Th/8nKBI5sxX5ykpQXKS7qtUuTIlKJvUL2kkfaer36TcpQbK5ZIa4akL1PkE4RUS6rzPvU1SNdxNU+oTE+k3gjd0GCuPEtEA0RXShEfRUIzUbeCKXp8H5j/mP+a//7kF1XOpfO15NRH5/cpArymqr393zlPX37ugOjjiM41MlddgCbx26DAj5jPrMVUMGWVjU4/Jq+QrbA+z6mt7zVD11PXtiup3PZRZ7yFnLp6hin3VT6rM+pqhzKwsgns9S8cK7o0LB1sRK8I+Jq8KxUXkc/RCGQL2fXqtGqXM6kn5WxRVQL/boa0IkBSf+Yz+KW5envBq8VybXiuUrQOjeVlBz9KZvSSwDszIilOzPLB4TfQVobiImL6kwnrG8+Rlg9cFvYfzvUopiwN4eZvgWxRljMes383w+ZpOVTqWFfU1jFdj1RRi+SdaQNWj6NEcOjx4BbwGZ8j5SqFOK4bHTr2VvhrUm9FqtgLe00TYr2hVlk90I9lsCrvrgk9lF7H7RJlh4olRcfKUnKcTljkVjs9qz5JKT0bqzdbP1jc0b6P81Ax73w4qktnKLyhtn2eJrXtaWTb76VPONzx+DThPi1im/YmhHjOvmtfPR+DbHntTJbQFoADG/S+J1NmU1fRQfgMo24oKAEa3AVUFjOsMmwKqb6AK/GxaeeQYAFo1JCBJcJWDTWMeyqjijX5J9t5UESMJ/AzIkkrX7rIxFL5psBS0cmuaiZtQ+RF9z9hltdxRI2xKt1Kqt+kERTyVATKlsLbjLIHijZE3myWX1IjZnjdGDmZX/Ko1/WDmZEqZZmf1A2CWCS0da2u6/4wFVY6nNzX8+9ka/f9GR5VZmVnN89T7bhbzGvCOBviY+dxPvbpGJ1nV2FXE42tbx6dvLDODGXkYTc3BNTwYADNgwGzocM0wldQrwGrcz6PBhtw9pGVmCzZcsOvCHD6i3vvpl+PcYjntqoBZawtmLOmtZXgr8RoYyIFmVipqpA4HAMDQhCEwsuiV6k09ojs+vfdUxnYVjXIAnkwZsC0z2iEwZr31jffeDN+CffrLZ7X13tMP69MPDPksPirggirLQMtq2aefRhn/qBSn9xCBiLRsRcD3EFgR2+qDttLQELYCQwoilu8hwub7kETA2orgXhqv85AGwc3p/2IcEjYNo/5ldOIjBdunPLxkHn2a9dFxlf8U9SVLRWs+7nOwvO1mMr7ro36dwyaFISBY9rMi6l5yueszyzLvb2E3x64bpZDd/t2yD/sg2M4e80Mf9UViZd4vi+1SWL8U1XtoD+7bddM4lG3yqetOi36aBkFwxLrPHp/62W3UHZD20THDgKQZOn0Igm2/GfUshz4fmpyA7H0X7WdcoC1V1ueB8YVMjeuOC/y2j8J1WKz6H4PtZNeXcdYXyz76lKyO0a8pBsyS0Bj9dTeXWRh1oyfAThddFC5rANU+6tfH0ZhPo9koCufj26hbMzBe9lH/MrVy0y/LPjpmj7MgCMId0mMX9ev8cR0EwQsuw6m11bHr9ilsMjqlo6hvAIlyo5J3xK9RDfLJTb6pVpCbalMDyWS3q6oV/nW/sqy6UzQAqwmvqs3Y5ZNqCow3GSvGk1U7n+QA5lUCA/bwkYiEg88FVr6/pkZ7eSzJ5Yn2adnTH1H10dvrXbBn7qgvy6Dfn6ojVXFFhawpFfCIdjjRh+qWTnwdUlEVdBzbgoJTeksfpKJCYA+Uxk/UN+meSplGFDaTJupnST2i36o1hcnjhIIvIM3g2bkVvk87UkmusAIruG8fGw1JH5TpiWief4Fp8YxPtIlL2p7SE/3Ql2+DsHYFTZMTFfNsTS+bJOzZvdA0eUdFJpugK9Mj/VcIiRKprKjP0waifZ4wbsQo/04iKHBH76pDRKe6CfrZOKN1PP8lCGdr2odF2geZGx68p0yQRB2ndEyfekrdKgqW1SW9Y0nN/nY576m7m3XdaJQuaCk87OQh6KYuIwpnb4M+z+9oMeeU+ll1SZe4OdJ6xWGwcildmf1MOoru0j1VlkuEsO4KitjdcQW9KNuM+mNE2/RttKXL2jW0v4qCYIkW7Q/75yGyoJduS5+yUURR5VYU1pAbenFLOrWuovD6W5gBa60PiYjj7OIE7UFF8r93Y7/FFtfRfuxk2v0RvO/yeLU9Ml9FN04WtEacUSETKtxN8ALJfvwgM7qzTg40+UyNOClpKWsqWodjlHPfj6Xu+zrv9hzHOAW7pOsS5/IuZLekf4rjeELrxw0VcPC5kA39CCs9iGxs4lfIGlsmO3zWSStrqpygoE3c0CzGwhBY43G33d44+UCbuKBV/HlY4JxmH3nI0KSkNOawu45ju2jNJJHfZxR+GOV9tsYXLapyS3shv+Le0CWczIOT5NERjxndxfMgTFy7p0biipp4gKtoKVJRGV/S7mM8XVd4v13FsVtSyl2fO0n6bsx/sIJK4oYu2dVdiBk1cRzPg4VkNMrd45Kyjxk1H3kfJHHsTkHi1jSP61G0G/fd1Ll6FEzzflvJ4w0Vgixi8UdSUEQnOJZuoh+L/KYL9hO+o1k8o+Aw3v1AG5GSbnbBkILCSV1QtB5nwZ530WjAigrHl5TWfbASNl5H3ovKCitvBW3zGmb9III7oaVjcNhjvI92bjqkGwFYUyWS0QkDLE7ba1fSxF0GU2HUWRL2Ndjm0WnAxLWyCBK3pEk8p3U8ofU0y6YHmuTRnoGkC3M72s5vspvnKHQbahz7vVpSbjnS8r1KEg7+4vw3rL++c9onrA+ojVpdSRtXUiMeNg4+QQ1xB3oSmdBb58UtafY4p9Mqy6bL1oeSt0nESU0uivqca/tGYGVDdo20FKkN1cDWLikVBkdH2UWj2hnQAuBTsBKZUDlE8n7P9kNw83hLxzKrY5dE4U2Wrf6Jig2tpW3lGI153yVWUkod0P/a/hRMVlmW9Vc2o6VjaSiNN1QK75J8la0GSEETNyAZioHdjKr6KmTYJPggqJPkzkqKWEdk000wYrGfaOJcRsu4oVtrXUafZPrL2q08ZE5W9JMIH4Nd4kVs0vefqJH2W/gU30evDtrOi46s5bFuS5ZUCqPuz2DPPnwBQyANVdj3XIfRzvI9xHBTjPuRIdjaBSUuC9bujrI4pe02CIIoSK+7fW2CkIYu22CEDZW+05+6XlmFjm23/jrPRq5WxjTaD5AzqQR1N2Lg9ssRP3mzuAbt2UpXZ1/6SJzViheaWMb4LD4NOG5vrFT0doiVNa3q6AS7KyKifsnP28BT7YxKqGI0JJpiiID7sD2Z/diGnNFBhnx2Q1Avu2iotPNhQ40wy4Rm46swh90F73xxb/QVhXJOjPZTkFmZG5S2BYfh9Q97fNwFJyu2oMTaKd0K6h+CqR8sk4AKQB+M3fc7z7Ob3A2lGeDZhV8GJNuXaXAnfNwmlnnAVdgaHMUHi2OfhCN2FZV8fX09rnka+dEl+fV1Xtfiw5c65bzOeXa8uymSzoflV8xhQEpFng+z+DSXqu92332P+H5cBU2EUQ+Ybr8MNqNyQqmwkzy9C2k9Dd7nRtpo6LMd7s7BD9up2Q94KYc4PlGxyesF+TGnxg2oKD0naUr7KXj2soS2cgxLeplUZRDOEiloOqQYcMaOJemolPbbnmkxf8pFLmRUOMi478E/RMmXAP/QfwgykTVNHNz8tLkMVg4uo4Uv8hQU1Ihk9EFE4vTdLjHI+3Asl3QTW1cvlm7ylTAhWqG43mAU5QJ3oE2cUuPOT0PpmTSUPibBHs7G1fMyQjV/mu/Ho2xnVDq4XXCSpPsygPfHRZfI9Yc1XJxHIfc9P9o4+6la0QeBcNidCdY0j61ri4P4MacyzugkIo8vtHNLD9PgBAu3pvm3FZ9ti6QbLfu+67pgG6XxWeyiEUN2VAj4FPRRJqzGzvdRYot9DCqsThSyW9DN+WllEtCJYbOgf8aqp92cRiu7CoO53HoBrnvqEqBd0DJHFXS5Sb2Pdu45COfIL6l8PAcKowDLjH6rx8tgQEYHOTfxhMqziDf0Wewd/ZZg3rXekuRS8/9GU8Z112WY7+lk875Lzg9LQ/Q7RdAeacao6FIaejfFKqS5PQYppu+oP5vVtt9wXtCd8yOjQnbbfiXJMgim8YxmjkfRVD7QocaMjpx8S7EWcRUtwczIgh9qyBlM8y5kJ80AaeiQBfsa3zngdhVSFx27kOUtBR8yCpmDfkBBqcicfrAVLaCoQ2qE2VYR9UGUCt4GgSnwQumA95TY1ha0sC0jX1B0RfuVm/7Sj6G4o+1v+KeO+ihYwlW0fCU8KPg3iqL+SJvHXUf9fEGTuKR/GmAE9UJp/Ilu3DSifrqh0tVdxG7hMcmwpm1PfVpEFzTlbc8koqvocnsCbmm7nNNtvKH3A/aUifwTLT8uqRS2uy3Nhe3NnrqORlPhpac9drqnbnvq+5r7X6ZWZjRz4KhjmfR0FdEdu2daDyhpKUlP3TwNgj7qO5q654D22Z5Wkr+j6IqOidsF3RA6fsPzy3HAe5oPOcQz2i7mTdQnFoCs6MiQhT8rd0mHeRFREc+prx+Xxh8nFjzJNkmbSR72DJ40VZ5OmNMJv0H2lAB5OkfylEEVz2kOVZZd2qQ7y8AqbdKpoB0KKuZpDcX0aTeUA89nTZULxsMSFOOq2UCStJllFkieVtCHV8CDgjfNU97QRrCazZIsTex0WLRPUJI95UA2m+XXT1NwVTEMk5vZLBFksya9dm+THgYP2mKaNs95dATyqsnyp0yu03n7hjdpbiQpq6cplLFJGcrIq1kzGXvbsxMw76pmkoRh3U7SMbAbdshIRq6rZpYByJ8yo3UrYDprdmLkdPM0BrLZU75Jc1jezGaT2qKuKgbG6RzSozfcdJcDFlFm5Tkq44ZOx210WoFV2a6GUnYd7QZEawGfuqnsLrfBaRbd4RRVgvwYbYTLxIKnDBHHfc9vIE4gwizCytbh/h4iDCesyiK4V1NTwKqtFRHrEVQW435wYK+cgM3HcOk9xAmbMq3ew4FfDQ8KEXksB7B1AnFgDIv2CcqonsUJ4GDk52fCrZiJeVg8KFsRVw94A3EWzjJE2GyHkSQczPa0ZvBntQfjq3cu/pjRhxYiuDc0zcB8+2HZbB1bcWBD3ZubYXtWP+8hwvT2oMp5zcp1ztxyXktDJfKc7dBUW0OqrXNm5rxuTT9wnjPymvMcYHA+/g6xQz6dxl/EUGZW5qGIPhgC6qNhehjjyuwVYa+g7/1YPbX4eHs01CPzYMq8Goy6xMA3In3wC8rvYQK+t59RtsS2JZcyM8bBET+fuz2ebpwRvF970F7SYbcJaS6ejvgEdlY/vKmxqk+4PpX69dOvgTTmExWgxnmqCgxYCgBTVP2k/BUAQAEfn+8QN8d7OdXhkMH2YIy+Gjamk5+B11JV/QTO0lI/uzNsVfUMmVf3OGtVD/4R+QblLc/PQvWssjFumKmqGEcn/Hzu9vhn6l+kmp7IF1EQ9BXYI3hGt86IT/3OUx8Xf0X/cmnurMaxNHSwraqeJeWv4GmFfxMeEg9mE5OE9RvySw4q8l0N/Zp+yUHl3atrjzI4me5yi9etvd+IaFY2Ozmlr6Ip31kAvjm5h8X915dXeRVQBqwddvc7Cg/KEOjDd0QejLf+viF5UGXVh68pD8r6Sqs034r/zsKDaqsP3xl5UP0m/freg4i+NpDUZZ83Mb9TcJ55+o6Eqdhd/3ccHv63/zD/NfdkkuegQTM442EdU4aLHtECyRHqcaqGXrBG4h1UxmqOS5HcOV+dHeEVSqmc8NEcl7KZk6GRNN7nLsmxT/jIOF2ZuZEiyVyjTpO+keikv6vK0vSNlE3mwzTyerwDfOevh03HO6ZubJzwgTERSmt9PBpP9kE5tS1nJ8CyZcx/zH/Mf8x/rH+5bONs0MwGZN33vaXNGXjL1L2h2Kb3ta+ED+cb42dvCse3doJdgr5uQF41tQCEY03xAdkjzjei55Cc/vHRc1A+63sDLHvlI63gqFxlsCadwZzvApR57Oo5SwHKRLh+I/izYYrUbcYESyw+bSDzH/Mf898mAalx9HA+xXPo6MUnqmwcKev5JJX1LIuLTlQBAHoP4P48lgpA/S0uOtF/RJ7k44RRJzXOgXKe1K16Ldo8YdWLTPRnlBGt72gV/1Mw83wNJMNVje0f8UKVsJoaLO+iBEO5i0qUsQvC56Tc7+KUhjAzD2U9Aj/j8/7Zsukz4BQkMMwuLHETOnwUeYQYMEQErGoICO5ZHsFWABHbejDMLirJCjqlyaoauyGQpHqqEmkZSZWmWQvOqsTepPUuTVfAEGuz9AYXkyjaIgi23XNBq3iAwywKuqBLYec9ddvglrGgKr6lRd9RVMkQV1G/woXlDhXdQgqaDokndLxBFgYZ9tHcXh9pLgVN4jvqUjxvQ5bTdjwJ9onwhSVS0Vq8uNM2/+g+rn4skq7LOOZpDpPKAaft1J2icnu8lgvMU+wA6yHNr7bvLi8/nKiv10Sj9QTiYSXgBWWP74iodKwPF6NcR92nxWLxsi4ZVdERHafWJDOYDzjto+zCdMd9z7E4uR6Pp+ySm4Ju5SxOQf5MR8bFKLP4/9OTiEvp84o+SPxxRcU5uJY1fZZWLyxJqRiyGvKYhFRslkG/w4leNmkfbOSFqnht8Inm8b9b57qT/rFpJsklVJp0B8hdN32sulSQFH3ULaYWybqPouMEuOs2j8tuZcFFl7kd/zFXuOds8xouYJKnzDUzc82tIWBR1zUsM8B1zRbqVVKvgeTi00dxCZYbZipg3H6vUFV+A+ANq7ICAJtKgOqD16BpU0FzLyjG7YO9IaGqhslQwaukA/uzDMB5l4PMf8x/zH8tRIHoSSuA85RSmj9O+OhVD1yp3zzvypf70BzmP+Y/5j/mPxYEHw==)

The figure shows the flow of the use case execution:

1. Identify scenes from a video stream coming through a camera source.
2. Compose semantic segmentation and video stream using qtivcomposer.
3. Encode the stream as H.264 bit stream and multiplex the stream in a MP4
                container.

The table provides the sequential processing stages of the pipeline execution:

| Process | Description |
| --- | --- |
| [qtiqmmfsrc](https://docs.qualcomm.com/doc/80-70014-50/topic/qtiqmmfsrc.html) | <ol class="ol" id="single-camera-stream-with-image-segmentation-and-encode__ol_f5k_g5n_vbc"><br>                                <li class="li">Collects the video stream (source) and creates two copies of the<br>                                        source:<ul class="ul" id="single-camera-stream-with-image-segmentation-and-encode__ul_n44_nwl_vbc"><br>                                        <li class="li">One stream is sent to the qtivcomposer plugin to retain<br>                                            the video stream.</li><br><br>                                        <li class="li">The other stream is sent to ML inferencing branch in the<br>                                            pipeline.</li><br><br>                                    </ul><br></li><br><br>                            </ol> |
| **Preprocessing** | **Preprocessing** |
| [qtimlvconverter](https://docs.qualcomm.com/doc/80-70014-50/topic/qtimlvconverter.html) | <ol class="ol" id="single-camera-stream-with-image-segmentation-and-encode__ol_xsf_q5l_vbc"><br>                                <li class="li">Receives the video stream on its sink pad.</li><br><br>                                <li class="li">Performs preprocessing:<ul class="ul" id="single-camera-stream-with-image-segmentation-and-encode__ul_ff2_twl_vbc"><br>                                        <li class="li">Color conversion</li><br><br>                                        <li class="li">Scaling down/up</li><br><br>                                        <li class="li">Normalization on the stream data when model expects<br>                                            floating point values as input</li><br><br>                                    </ul><br></li><br><br>                                <li class="li">Converts the video stream to a tensor stream on its source<br>                                        pad.<p class="p">The segmentation model uses this tensor stream for<br>                                        inferencing.</p><br></li><br><br>                            </ol> |
| **Inferencing** | **Inferencing** |
| [qtimltflite](https://docs.qualcomm.com/doc/80-70014-50/topic/qtimltflite.html) | <ol class="ol" id="single-camera-stream-with-image-segmentation-and-encode__ol_lfr_35n_vbc"><br>                                <li class="li">Loads the segmentation model.</li><br><br>                                <li class="li">Modifies the graph for the chosen delegate.</li><br><br>                                <li class="li">Receives the tensor stream on its sinkpad.</li><br><br>                                <li class="li">Executes the inference and produces tensor stream with the<br>                                    segmentation results on its source pad.</li><br><br>                            </ol> |
| **Postprocessing** | **Postprocessing** |
| [qtimlvsegmentation](https://docs.qualcomm.com/doc/80-70014-50/topic/qtimlvsegmentation.html) | <ol class="ol" id="single-camera-stream-with-image-segmentation-and-encode__ol_mtr_k5n_vbc"><br>                                <li class="li">Receives the inference tensors on its sinkpad.</li><br><br>                                <li class="li">Converts the inference tensors into video formats that the<br>                                    multimedia plugins can process later.</li><br><br>                                <li class="li">Produces the semantic segmentations for the frame.</li><br><br>                                <li class="li">Loads the corresponding modules for the segmentation<br>                                        models.<p class="p">In this use case, qtimlvsegmentation does the<br>                                        following: </p><ol class="ol" type="a" id="single-camera-stream-with-image-segmentation-and-encode__ol_ntr_k5n_vbc"><br>                                        <li class="li">Loads the deeplab-argmax submodule.</li><br><br>                                        <li class="li">Produces video frames with segmentation masks.</li><br><br>                                        <li class="li">Sends them to sinkpad of qtivcomposer.</li><br><br>                                    </ol><br><br>                                </li><br><br>                            </ol> |
| [qtivcomposer](https://docs.qualcomm.com/doc/80-70014-50/topic/qtivcomposer.html) | <ol class="ol" id="single-camera-stream-with-image-segmentation-and-encode__ol_nmc_lxl_vbc"><br>                                <li class="li">Receives the original video stream with segmentation mask on its<br>                                    sinkpads. </li><br><br>                                <li class="li">Produces on its sourcepad GST buffers with contents composed of<br>                                    video streams from its sinkpads.</li><br><br>                            </ol> |
| [v4l2h264enc](https://docs.qualcomm.com/doc/80-70014-50/topic/v4l2h264enc.html) | <ol class="ol" id="single-camera-stream-with-image-segmentation-and-encode__ol_wsc_bsn_vbc"><br>                                <li class="li">Applies parameters to each frame of the video stream its<br>                                    receiving on its sinkpad.</li><br><br>                                <li class="li">Encodes it into bitstream and sends it over its sourcepad.</li><br><br>                            </ol> |
| h264parse | Adds additional information corresponding to the bitstream to<br>                            GStreamer buffer meta. |
| mp4mux | Receives these buffers and creates containers format specification<br>                            buffers. |
| **Output** | **Output** |
| Filesink | Stores the resulting stream in a /opt/video.mp4<br>                            file. |
| Playback | Use the following command to pull video.mp4 from<br>                            the host machine and play it on a media player<br>                                    application:<br>`scp root@<IP address of target<br>                                    device>:/opt/ <destination directory>` |

**Parent Topic:** [TensorFlow Lite use cases](https://docs.qualcomm.com/doc/80-70014-50/topic/tensorflow-lite-use-cases.html)

Last Published: Oct 27, 2025

[Previous Topic
Image segmentation and display with TFLite](https://docs.qualcomm.com/bundle/publicresource/80-70014-50/topics/single-camera-stream-with-image-segmentation-and-display.md) [Next Topic
Pose estimation and display with TFLite](https://docs.qualcomm.com/bundle/publicresource/80-70014-50/topics/single-camera-stream-with-pose-estimation-and-display.md)