Konrad Lis
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Konrad Lis has a total of 26 co-authors. Here are the top co-authors by number of shared publications:
- Tomasz Kryjak
AGH University of Krakow
7 shared publications - Marek Gorgon
AGH University of Krakow
4 shared publications - Joanna Stanisz
AGH Akademia Górniczo-Hutnicza
2 shared publications - Adam Glowacz
Department of Automatic Control and Robotics, Faculty of Electrical Engineering, Automatics, Computer Science and Biomedical Engineering, AGH University of Science and Technology, Al. A. Mickiewicza 30, 30-059 Krakow, Poland
1 shared publication - Aleksander Kostuch
AGH University of Krakow
1 shared publication - Andrzej Brodzicki
Department of Automatic Control and Robotics, AGH University of Science and Technology, 30-059 Krakow, Poland
1 shared publication
Latest Publications
- 1 April 2025
- Published by Springer Nature
- p. 28-40
LiFT: Lightweight, FPGA-Tailored 3D Object Detection Based on LiDAR Data
- 1 April 2025
- Published by Springer Nature
- p. 28-40
LiFT: Lightweight, FPGA-Tailored 3D Object Detection Based on LiDAR Data
- Open Access
LiFT: Lightweight, FPGA-tailored 3D object detection based on LiDAR data
This paper presents LiFT, a lightweight, fully quantized 3D object detection algorithm for LiDAR data, optimized for real-time inference on FPGA platforms. Through an in-depth analysis of FPGA-specific limitations, we identify a set of FPGA-induced constraints that shape the algorithm's design. These include a computational complexity limit of 30 GMACs (billion multiply-accumulate operations), INT8 quantization for weights and activations, 2D cell-based processing instead of 3D voxels, and minimal use of skip connections. To meet these constraints while maximizing performance, LiFT combines novel mechanisms with state-of-the-art techniques such as reparameterizable convolutions and fully sparse architecture. Key innovations include the Dual-bound Pillar Feature Net, which boosts performance without increasing complexity, and an efficient scheme for INT8 quantization of input features. With a computational cost of just 20.73 GMACs, LiFT stands out as one of the few algorithms targeting minimal-complexity 3D object detection. Among comparable methods, LiFT ranks first, achieving an mAP of 51.84% and an NDS of 61.01% on the challenging NuScenes validation dataset. The code will be available at https://github.com/vision-agh/lift.
Open AccessLiFT: Lightweight, FPGA-tailored 3D object detection based on LiDAR data
This paper presents LiFT, a lightweight, fully quantized 3D object detection algorithm for LiDAR data, optimized for real-time inference on FPGA platforms. Through an in-depth analysis of FPGA-specific limitations, we identify a set of FPGA-induced constraints that shape the algorithm's design. These include a computational complexity limit of 30 GMACs (billion multiply-accumulate operations), INT8 quantization for weights and activations, 2D cell-based processing instead of 3D voxels, and minimal use of skip connections. To meet these constraints while maximizing performance, LiFT combines novel mechanisms with state-of-the-art techniques such as reparameterizable convolutions and fully sparse architecture. Key innovations include the Dual-bound Pillar Feature Net, which boosts performance without increasing complexity, and an efficient scheme for INT8 quantization of input features. With a computational cost of just 20.73 GMACs, LiFT stands out as one of the few algorithms targeting minimal-complexity 3D object detection. Among comparable methods, LiFT ranks first, achieving an mAP of 51.84% and an NDS of 61.01% on the challenging NuScenes validation dataset. The code will be available at https://github.com/vision-agh/lift.
- Open Access
- 10 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in TechRxiv
Comparative study of subset selection methods for rapid prototyping of 3D object detection algorithms
Object detection in 3D is a crucial aspect in the context of autonomous vehicles and drones. However, prototyping detection algorithms is time-consuming and costly in terms of energy and environmental impact. To address these challenges, one can check the effectiveness of different models by training on a subset of the original training set. In this paper, we present a comparison of three algorithms for selecting such a subset - random sampling, random per class sampling, and our proposed MONSPeC (Maximum Object Number Sampling per Class). We provide empirical evidence for the superior effectiveness of random per class sampling and MONSPeC over basic random sampling. By replacing random sampling with one of the more efficient algorithms, the results obtained on the subset are more likely to transfer to the results on the entire dataset. The code is available at: https://github.com/vision-agh/monspec.
- 10 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in TechRxiv
Open AccessComparative study of subset selection methods for rapid prototyping of 3D object detection algorithms
Object detection in 3D is a crucial aspect in the context of autonomous vehicles and drones. However, prototyping detection algorithms is time-consuming and costly in terms of energy and environmental impact. To address these challenges, one can check the effectiveness of different models by training on a subset of the original training set. In this paper, we present a comparison of three algorithms for selecting such a subset - random sampling, random per class sampling, and our proposed MONSPeC (Maximum Object Number Sampling per Class). We provide empirical evidence for the superior effectiveness of random per class sampling and MONSPeC over basic random sampling. By replacing random sampling with one of the more efficient algorithms, the results obtained on the subset are more likely to transfer to the results on the entire dataset. The code is available at: https://github.com/vision-agh/monspec.
- Open Access
Comparative study of subset selection methods for rapid prototyping of 3D object detection algorithms
Object detection in 3D is a crucial aspect in the context of autonomous vehicles and drones. However, prototyping detection algorithms is time-consuming and costly in terms of energy and environmental impact. To address these challenges, one can check the effectiveness of different models by training on a subset of the original training set. In this paper, we present a comparison of three algorithms for selecting such a subset - random sampling, random per class sampling, and our proposed MONSPeC (Maximum Object Number Sampling per Class). We provide empirical evidence for the superior effectiveness of random per class sampling and MONSPeC over basic random sampling. By replacing random sampling with one of the more efficient algorithms, the results obtained on the subset are more likely to transfer to the results on the entire dataset. The code is available at: https://github.com/vision-agh/monspec.
Open AccessComparative study of subset selection methods for rapid prototyping of 3D object detection algorithms
Object detection in 3D is a crucial aspect in the context of autonomous vehicles and drones. However, prototyping detection algorithms is time-consuming and costly in terms of energy and environmental impact. To address these challenges, one can check the effectiveness of different models by training on a subset of the original training set. In this paper, we present a comparison of three algorithms for selecting such a subset - random sampling, random per class sampling, and our proposed MONSPeC (Maximum Object Number Sampling per Class). We provide empirical evidence for the superior effectiveness of random per class sampling and MONSPeC over basic random sampling. By replacing random sampling with one of the more efficient algorithms, the results obtained on the subset are more likely to transfer to the results on the entire dataset. The code is available at: https://github.com/vision-agh/monspec.
- 11 February 2023
- Published by Springer Nature
- p. 99-119
PointPillars Backbone Type Selection for Fast and Accurate LiDAR Object Detection
- 11 February 2023
- Published by Springer Nature
- p. 99-119
PointPillars Backbone Type Selection for Fast and Accurate LiDAR Object Detection
- Open Access
- 5 October 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in TechRxiv
PointPillars Backbone Type Selection For Fast and Accurate LiDAR Object Detection
3D object detection from LiDAR sensor data is an important topic in the context of autonomous cars and drones. In this paper, we present the results of experiments on the impact of backbone selection of a deep convolutional neural network on detection accuracy and computation speed. We chose the PointPillars network, which is characterised by a simple architecture, high speed, and modularity that allows for easy expansion. During the experiments, we paid particular attention to the change in detection efficiency (measured by the mAP metric) and the total number of multiply-addition operations needed to process one point cloud. We tested 10 different convolutional neural network architectures that are widely used in image-based detection problems. For a backbone like MobilenetV1, we obtained an almost 4x speedup at the cost of a 1.13% decrease in mAP. On the other hand, for CSPDarknet we got an acceleration of more than 1.5x at an increase in mAP of 0.33%. We have thus demonstrated that it is possible to significantly speed up a 3D object detector in LiDAR point clouds with a small decrease in detection efficiency. This result can be used when PointPillars or similar algorithms are implemented in embedded systems, including SoC FPGAs. The code is available at https://github.com/vision-agh/pointpillars\_backbone.
- 5 October 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in TechRxiv
Open AccessPointPillars Backbone Type Selection For Fast and Accurate LiDAR Object Detection
3D object detection from LiDAR sensor data is an important topic in the context of autonomous cars and drones. In this paper, we present the results of experiments on the impact of backbone selection of a deep convolutional neural network on detection accuracy and computation speed. We chose the PointPillars network, which is characterised by a simple architecture, high speed, and modularity that allows for easy expansion. During the experiments, we paid particular attention to the change in detection efficiency (measured by the mAP metric) and the total number of multiply-addition operations needed to process one point cloud. We tested 10 different convolutional neural network architectures that are widely used in image-based detection problems. For a backbone like MobilenetV1, we obtained an almost 4x speedup at the cost of a 1.13% decrease in mAP. On the other hand, for CSPDarknet we got an acceleration of more than 1.5x at an increase in mAP of 0.33%. We have thus demonstrated that it is possible to significantly speed up a 3D object detector in LiDAR point clouds with a small decrease in detection efficiency. This result can be used when PointPillars or similar algorithms are implemented in embedded systems, including SoC FPGAs. The code is available at https://github.com/vision-agh/pointpillars\_backbone.
- 1 January 2022
- Published by AGH University Press
- p. 17-31
Systemy wizyjne w zastosowaniach przemysłowych
- 1 January 2022
- Published by AGH University Press
- p. 17-31
Systemy wizyjne w zastosowaniach przemysłowych
- Open Access
- 27 December 2021
- Published by Springer Nature in Journal of Signal Processing Systems
- Vol. 94 (7) , 659-674
Implementation of the PointPillars Network for 3D Object Detection in Reprogrammable Heterogeneous Devices Using FINN
In this paper, we present a hardware-software implementation of a deep neural network for object detection based on a point cloud obtained by a LiDAR sensor. The PointPillars network was used in the research, as it is a reasonable compromise between detection accuracy and calculation complexity. The Brevitas / PyTorch tools were used for network quantisation (described in our previous paper) and the FINN tool for hardware implementation in the reprogrammable Zynq UltraScale+ MPSoC device. The obtained results show that quite a significant computation precision limitation along with a few network architecture simplifications allows the solution to be implemented on a heterogeneous embedded platform with maximum 19% AP loss in 3D, maximum 8% AP loss in BEV and execution time 375ms (the FPGA part takes 262ms). We have also compared our solution in terms of inference speed with a Vitis AI implementation proposed by Xilinx (19 Hz frame rate). Especially, we have thoroughly investigated the fundamental causes of differences in the frame rate of both solutions. The code is available at https://github.com/vision-agh/pp-finn.
- 27 December 2021
- Published by Springer Nature in Journal of Signal Processing Systems
- Vol. 94 (7) , 659-674
Open AccessImplementation of the PointPillars Network for 3D Object Detection in Reprogrammable Heterogeneous Devices Using FINN
In this paper, we present a hardware-software implementation of a deep neural network for object detection based on a point cloud obtained by a LiDAR sensor. The PointPillars network was used in the research, as it is a reasonable compromise between detection accuracy and calculation complexity. The Brevitas / PyTorch tools were used for network quantisation (described in our previous paper) and the FINN tool for hardware implementation in the reprogrammable Zynq UltraScale+ MPSoC device. The obtained results show that quite a significant computation precision limitation along with a few network architecture simplifications allows the solution to be implemented on a heterogeneous embedded platform with maximum 19% AP loss in 3D, maximum 8% AP loss in BEV and execution time 375ms (the FPGA part takes 262ms). We have also compared our solution in terms of inference speed with a Vitis AI implementation proposed by Xilinx (19 Hz frame rate). Especially, we have thoroughly investigated the fundamental causes of differences in the frame rate of both solutions. The code is available at https://github.com/vision-agh/pp-finn.
- Open Access
- 13 March 2008
- Published by American Physical Society (APS) in Physical Review B
- Vol. 77 (11) , 115320
Quantum dot defined in a two-dimensional electron gas at aheterojunction: Simulation of electrostatic potential and charging propertiesWe present a self-consistent Schrödinger–Poisson scheme for simulation of electrostatic quantum dots defined in gated two-dimensional electron gas formed at heterojunction. The computational method is applied to a quantitative description of transport properties studied experimentally by Elzermann et al. [Appl. Phys. Lett. 84, 4617 (2004)]. Our three-dimensional model describes the electrostatics of the entire device with a quantum dot that changes shape and floats inside a gated region when the applied voltages are varied. Our approach accounts for the metal electrodes of arbitrary geometry, includes magnetic field applied perpendicular to the growth direction, electron-electron correlation in the confined electron system, and its interaction with the electron reservoir surrounding the quantum dot. We calculate the electric field, the space charge distribution, and energies as well as wave functions of confined electrons to describe opening of two transport channels between the reservoir and the confined charge puddle. We determine the voltages for charging the dot with up to four electrons. The results are in qualitative and quantitative agreement with the experimental data.- 13 March 2008
- Published by American Physical Society (APS) in Physical Review B
- Vol. 77 (11) , 115320
Open AccessQuantum dot defined in a two-dimensional electron gas at aheterojunction: Simulation of electrostatic potential and charging propertiesWe present a self-consistent Schrödinger–Poisson scheme for simulation of electrostatic quantum dots defined in gated two-dimensional electron gas formed at heterojunction. The computational method is applied to a quantitative description of transport properties studied experimentally by Elzermann et al. [Appl. Phys. Lett. 84, 4617 (2004)]. Our three-dimensional model describes the electrostatics of the entire device with a quantum dot that changes shape and floats inside a gated region when the applied voltages are varied. Our approach accounts for the metal electrodes of arbitrary geometry, includes magnetic field applied perpendicular to the growth direction, electron-electron correlation in the confined electron system, and its interaction with the electron reservoir surrounding the quantum dot. We calculate the electric field, the space charge distribution, and energies as well as wave functions of confined electrons to describe opening of two transport channels between the reservoir and the confined charge puddle. We determine the voltages for charging the dot with up to four electrons. The results are in qualitative and quantitative agreement with the experimental data.