.. raw:: html

   <style> 
   body { background-color : #ffffcc; }
   h1, h2, h3, h4, h5, h6 { text-align : center; }
   .centered { text-align : center; vertical-align : middle; }
   .hidden { display: none }
   </style>


=======================================
CI1176 - Tópicos em Visão Computacional
=======================================

BCC, IBM \& PPGInf - Turma 2023-1
---------------------------------


**Professor:** `David Menotti <http://web.inf.ufpr.br/menotti>`__ `@inf.ufpr.br <to:menotti@inf.ufpr.br>`__ `@gmail.com  <to:menottid@gmail.com>`__

**Carga Horária:** 60 horas em aulas expositivas e colaborativas.

**Horário:** terças-feiras e quintas-feiras das 15h40 às 17h20

**Sala:** PA 02 (ter) & PC 18 (qui) (Presencial) / Google Meet (Digital)

------


Avaliação
------------

**80%** - quatro apresentações de pesos iguais

**20%** - frequência
  - 100% dos 20%: no máximo 3 ausências
  - 50% dos 20%: entre 5 e 4 ausências
  - 0% dos 20%: entre 7 e 6 ausências

------


Ementa
------

- Leitura, estudo e apresentação de artigos científicos da área publicados em veículos de impacto e relevância relacionados a diversos tópicos da área de visão computacional. 

------



Datas Importantes
-----------------

-  21/03/23 **Primeiro Encontro**

-  06/04/23 **Quinta-feira santa** (não haverá aula)

-  01/06/23 **Feira de Cursos e Profissões** (não haverá aula)

-  08/06/23 **Corpus Christi**

-  06/07/23 **Final**

------


Cronograma Tentativo
--------------------

.. include:: schedule.rst

------



Tópicos
-------

.. include:: topics.rst

----------


Referências
-----------

.. bibliography::

-----------


Referências por tópico
----------------------

- Background 1:                   ForwardForward :cite:p:`forwardforward:2022`, Backprop :cite:p:`backprop:1989`, LeNet :cite:p:`lenet:1998`, AlexNet :cite:p:`alexnet:2012`.

- Background 2:                   ResNet :cite:p:`resnet:2016`, LSTM :cite:p:`lstm:1997`, LSTM :cite:p:`lstm:2009`, Vision Transformers (ViT) :cite:p:`vit:2021`.

- Object Classification:          VGGNet :cite:p:`oc-vggnet:2015`, MobileNet :cite:p:`oc-mobilenet:2017`, EfficientNet :cite:p:`oc-efficientnet:2020`. 

- Object Detection:               Yolo :cite:p:`od-yolo:2016` & YoloV2 :cite:p:`od-yolo:2017`, Faster R-CNN :cite:p:`od-faster-rcnn:2017` , EfficientDet :cite:p:`od-efficientdet:2020`

- Image Segmentation:             Mask R-CNN :cite:p:`seg-maskrcnn:2018`, U-Net :cite:p:`seg-unet:2015`, IFT: :cite:p:`seg-ift-diff:2004`.

- Generative Models:              GAN :cite:p:`gm-gan:2014` & pix2pix :cite:p:`gm-pix2pix:2017`, CycleGAN: :cite:p:`gm-cyclegan:2017`, Stable Diffusion :cite:p:`gm-diffusion:2022`

- Vehicle Identification:         Efficient :cite:p:`alpr-efficient:2021`, Multi :cite:p:`alpr-multi:2020`, Bias :cite:p:`alpr-bias:2022`

- ALPR SuperResolution:           Realtime :cite:p:`alprsr-multitask:2019`, MPRNet :cite:p:`alprsr-mprnet:2021`, Attention Module & Pixel Shuffle :cite:p:`alprsr-combining:2022`.

- Face Detection:                 Viola-Jones: :cite:p:`fd-viola:2001`, RetinaFace :cite:p:`fd-retina:2020`,  MT-CNN :cite:p:`fn-mtcnn:2016`.

- Face Recognition:               Center loss :cite:p:`fr-centerloss:2016`, Arcface :cite:p:`fr-arcface:2019`, Partial FC :cite:p:`fr-partialfc:2022`.

- Face Rec. 3D:                   3D-PointFace :cite:p:`fr3d-pointface:2021`,  3D-PointCloudNet :cite:p:`fr3d-pointcloudnet:2021`, 3D-BERL :cite:p:`fr3d-3dberl:2022`

- Face Masked Rec.:               MaskInvArcface :cite:p:`mfr-synthetic:2022`, MaskInv-Hg :cite:p:`mfr-kd:2021`, Sota :cite:p:`mfr-sota:2022`.

- Face Recognition SR:            Iterative :cite:p:`frsr-iterative:2020`, SCTANet :cite:p:`frsr-sctanet:2023`, DIDNet :cite:p:`frsr-didnet:2021`.

- General Anti-Spoofing:          Fingerprint :cite:p:`gspf-fingerprint:2021`, Iris :cite:p:`gspf-iris:2018`, Voice :cite:p:`gspf-voice:2023`.
 
- Face Anti-Spoofing (passive):   FAS-CASIA :cite:p:`fas-casia:2012`, FAS-ViT :cite:p:`fas-vit:2021`, FAS-DC-CDN :cite:p:`fas-dc-cdn:2021`.

- Face Anti-Spoofing (active):    Distortion :cite:p:`fasa-distortion:2019`, Reflections :cite:p:`fasa-reflections:2015`, Sensor :cite:p:`fasa-sensor:2014`.

- Facial Expression:              Hybrid :cite:p:`fe-hybrid:2021`, Synthesis :cite:p:`fe-synthesis:2020`, Fusion :cite:p:`fe-fusion:2022`.

- Natural Language Processing:    Tesseract :cite:p:`docs-tesseract:2007`, Word2Vect :cite:p:`docs-word2vec1:2013` & :cite:p:`docs-word2vec2:2013`, Wordpiece :cite:p:`docs-wordpiece:2016`, GloVe :cite:p:`docs-glove:2014`, BERT :cite:p:`docs-bert:2018`, GPT :cite:p:`docs-gpt3:2020`.

- Document Recognition:           Intellix :cite:p:`er-intellix:2013`, GraphDoc :cite:p:`er-graphdoc:2022`, ERNIE-Layout :cite:p:`er-ernielayout:2022`.

- Action Recognition:             Trajectories :cite:p:`ar-trajectories:2013`, Kinetics :cite:p:`ar-kinetics:2017`, InternVideo :cite:p:`ar-internvideo:2022`.

- Reinforcement Learning:         DDPG :cite:p:`rl-ddpg:2020`, Pedestrian Trajectory :cite:p:`rl-pedestrian:2020`, Autonomous Driving :cite:p:`rl-driving:2020`.

- Zero Shot Learning:             Tell me :cite:p:`zsl-tellme:2021`, Global Semantic :cite:p:`zsl-global:2022`, Cezsar :cite:p:`zsl-cezsar:2023`.
 
- Gaze Estimation:                MPIIGaze :cite:p:`gaze-mpii:2019` , Gaze360 :cite:p:`gaze-360:2019` , L2CS-Net :cite:p:`gaze-l2csnet:2022`.

- Pose Estimation:                Hourglass :cite:p:`pose-hourglass:2016`, OpenPose :cite:p:`pose-openpose:2017`, Multi-Person with Transformers: :cite:p:`pose-transformers:2022`

- Object ReID (Person & Vehicle): Person in the Wild :cite:p:`reid-person:2017`, Person Ensemble :cite:p:`reid-personensemble:2020`, Vehicle Large :cite:p:`reid-vehiclelarge:2016`, Vehicle Read :cite:p:`reid-vehicleread:2021`.

- Biometrics:                     Iris :cite:p:`bio-irina:2017`, Fingerprint :cite:p:`bio-fingernet:2019`, Periocular :cite:p:`bio-ufpr-periocular:2022`

-----------


Frequência
----------

.. include:: students.rst

----------


.. include:: ../include/tracker.rst