About me

I’m currently a Postdoctoral Researcher at the Media Integration and Communication Center of the University of Florence. I received the PhD degree in Computer Engineering, Multimedia and Telecommunication from the University of Florence in 2012 with a thesis on “Wide area surveillance with rotating and zooming cameras”, under the supervision of Prof. Alberto Del Bimbo. I have published around 25 papers in the most prestigious journals and conferences in computer vision, multimedia and image processing. I gave a tutorial at a conference of international relevance and I have been part of the program committee of several conferences and workshops of international relevance. I am involved in research collaborations with other research centers, both national and international and I also work in tight collaboration with industrial partners, with particular emphasis on technology and knowledge transfer.

My research interests focus on computer vision, pattern recognition and machine learning for person detection and tracking, person re-identification and face recognition. Visit my profiles on Google Scholar and Scopus.

View Giuseppe Lisanti's profile on LinkedInFollow Giuseppe Lisanti on Twitter
Posted in Uncategorized | Comments Off on About me

WHOS descriptors for the Market-1501 dataset

We released the Weighted Histogram of Overlapping Stripes descriptors for the Market-1501 dataset. Both versions, HS+RGB+Lab+HOG+LBP and HS+RGB+HOG have been extracted.

Descriptors are available here.

Related articles:

  1. Lisanti et al., “Multi channel-kernel canonical correlation analysis for cross-view person re-id”, ACM TOMM, 2017.
  2. Lisanti et al., “Matching people across camera views using kernel canonical correlation analysis”, ICDSC, 2014.
  3. Karaman et al., “Leveraging local neighborhood topology for large scale person re-identification”, PR, 2014.
Posted in Latest News | Comments Off on WHOS descriptors for the Market-1501 dataset

International Workshop on Cross-domain Human Identification

I’m organizing the first edition of the “International Workshop on Cross-domain Human Identification”. Visit our homepage.

The aim of this workshop is to bring together leading researchers working on automatic human recognition to advocate and promote new research directions to video-surveillance as well as other, less obvious, domains such as entertainment, social network analysis, privacy preservation, customer behavior analysis, de-identification methods.

Posted in Latest News | Comments Off on International Workshop on Cross-domain Human Identification

MCK-CCA for Person Re-Identification

Our paper “Multi channel-kernel canonical correlation analysis for cross-view person re-identification” has been accepted for publication on ACM Transactions on Multimedia Computing, Communications, and Applications (TOMM).

The article is available at ACM TOMM

The source code is available on GitHub

Posted in Latest News | Comments Off on MCK-CCA for Person Re-Identification

Frontalized Faces in the Wild

We released the Frontalized Faces in the Wild dataset, which contains the frontalized version of the images collected in the famous and publicly available Labeled Faces in The Wild (LFW) dataset.

The dataset is available here.

Related articles:

  1. Ferrari et al., “Effective 3D based frontalization for unconstrained face recognition”, ICPR, 2016.
  2. Ferrari et al., “Dictionary Learning based 3D Morphable Model Construction for Face Recognition with Varying Expression and Pose”, 3DV, 2015.
Posted in Latest News | Comments Off on Frontalized Faces in the Wild