Geon-Woo Kim

On the Challenges of Applying Selective Encryption on Region-of-Interest in H.264 Video Coding

Selective encryption of video data is not a new idea and various techniques proposed in previous literatures can effectively protect sensitive data from unauthorized access. After the initial appearance of selective encryption, some principles have become well known and widely adopted. One of the principles is the standard format compliance, which means that unauthorized users can decode and process the encrypted video without knowing the content. In the context of ROI encryption, this allows unauthorized users to decode and view the video except for the encrypted region containing sensitive information. However, there are some challenges for providing format compliance in H.264 video coding when applying encryption on certain regions and it is very important to address this issue in some applications such as privacy masking in video surveillance system. In this work, we describe the challenges and introduce some approaches to address these. Due to the bulky size, applying standard ciphers to multimedia data tends to be inefficient for real-time processing scenarios. Therefore, selective encryption initially appeared as means to relieve computational cost and some principles for designing selective encryption scheme have become widely adopted. Those principles include security, time efficiency, format compliance, compression performance and error robustness. The format compliance means that the encryption process does not change the encoded bitstreams data format in order to support such direct operations as browsing, playing, cutting, copying and so on. In the past decade, several selective encryption algorithms supporting format compliance have been reported, most of which are based on MPEG and H.264 video codec. In the context of ROI encryption, format compliant encryption algorithm allows unauthorized users to decode and view the video except for the encrypted region containing sensitive information. In the application of video surveillance system, the problem of privacy invasion can be addressed by encrypting only a privacy sensitive region and, if the encryption algorithm guarantees format compliance, a person with a lower level of security clearance can only get the non- sensitive information, and all the privacy information can be entirely unperceivable to him/her. Although ROI encryption supporting format compliance is simply the application of established selective encryption algorithm on a given region, ROI encryption in currently prevailing video codec, such as H.264 and MPEG4, has a number of unique challenges. The challenges are all about confining the degradation of visual quality to the specified region and keeping the other parts intact. However, most of the consisting elements in the encoded video do have little meaning in itself and can reconstruct a valid data only when combined with other elements. When a consisting element outside a ROI is dependent upon the element inside a ROI and the related element is encrypted, the reconstructed value outside a ROI can be unexpected one and can introduce artifacts in decoded video. As every video codec specifies different encoding algorithm, the challenges are dependent on the employed codec. In case of MPEG4, two challenges are identified and the solutions are described in (1). Coding schemes for independent ROI encryption in scalable video coding are presented in (2).

Considerations on security model of home network

As various mobile sensing technologies, remote control and ubiquitous infrastructure are developing and expectations on quality of life are increasing, a lot of researches and developments on home network technologies and services are actively on going. Until now, we have focused on how to provide users with high-level home network services, while not many researches on home network security for guaranteeing safety are progressing. So, in this paper, we propose an access control model specific to home network that provides various kinds of users with home network services up to ones characteristics and features, and protects home network systems from illegal/unnecessary accesses or intrusions

Compact deep learned feature-based face recognition for Visual Internet of Things

The Visual Internet of Things has received much attention in recent years due to its ability to get the object location via image information of the scene, attach the visual label to the object, and then return information of scene objects to the network. In particular, face recognition is one of the most suitable means to Visual IoT because face feature is inherent label for human being. However, current state-of-the-art face recognition methods based on huge deep neural networks are difficult to apply in the embedded platform for Visual IoT due to the lack of computational resources. To solve this problem, we present compact deep neural network-based face recognition method for Visual Internet of Things. The proposed method has a low model complexity to operate in an embedded environment while using deep neural networks, which is strong against posture and illumination changes. We show competitive accuracy and performance results for the LFW verification benchmark and the collected mobile face recognition dataset. Additionally, we demonstrate that the implementation of the proposed system can be run in real time on the Android-based mobile embedded platform.

Security model for video surveillance system

Video surveillance system recognizes, keeps track of security threats of the real environment which threatens personal safety, and protects the individuals from them with visual devices gathering video information such as CCTVs and IP cameras. As it is widely deployed on open IP-based network, all security threats considered in the legacy IP-based applications might threaten the reliable applications of video surveillance, resulting in critical privacy infringement, misuse of video resource, increase of unexpected intelligent crime using unauthorized video access, and so on. So in this paper, we identify the security model for video surveillance system, ensuring reliability, safety and privacy protection.

De-identification of metering data for smart grid personal security in intelligent CCTV-based P2P cloud computing environment

Various security threats exist in the smart grid environment due to the fact that information and communication technology are grafted onto an existing power grid. In particular, smart metering data exposes a variety of information such as users’ life patterns and devices in use, and thereby serious infringement on personal information may occur. Therefore, we are in a situation where a de-identification algorithm suitable for metering data is required. Hence, this paper proposes a new de-identification method for metering data. The proposed method processes time information and numerical information as de-identification data, respectively, so that pattern information cannot be analyzed by the data. In addition, such a method has an advantage that a query such as a direct range search and aggregation processing in a database can be performed even in a de-identified state for statistical processing and availability.

Deep feature learning for person re-identification in a large-scale crowdsourced environment

Finding the same individual across cameras in disjoint views at different locations and times, which is known as person re-identification (re-id), is an important but difficult task in intelligent visual surveillance. However, to build a practical re-id system for large-scale and crowdsourced environments, the existing approaches are largely unsuitable because of their high model complexity. In this paper, we present a deep feature learning framework for automated large-scale person re-id with low computational cost and memory usage. The experimental results show that the proposed framework is comparable to the state-of-the-art methods while having low model complexity.

Prototype System Design for Large-Scale Person Re-identification

Identifying a person across cameras in disjoint views at different time and location has important applications in visual surveillance. However, it is difficult to apply existing methods to the development of large-scale person identification systems in practice due to underlying limitations such as high model complexity and batch learning with the labeled training data. In this paper, we propose a prototype system design for large-scale person re-identification that consists of two phases. In order to provide scalability and response within an acceptable time, and handle unlabeled data, we employ an agglomerative hierarchical clustering with simple matching and compact deep neural network for feature extraction.

Face Recognition for Mobile Self-authentication with Online Model Update

Face recognition system encounters complex change that varies over time, due to a limited control over the environment. So, the facial model of an individual tends to diverse from underlying distribution that collected during initial enrollment. However, new samples that are obtained each time people try to recognize or authenticate can be used to update and refine the models. In this paper, an efficient semi-supervised learning strategy is proposed to update the face recognition model. To maintain a high performance, we exploit a probability based update approach. Performance is assessed in terms of accuracy and equal error rate (EER). Experimental results illustrate that the proposed method effectively update the classifiers.

Privacy Preservation in SAT (Single Authentication Through)

In SAT (Single Authentication Through) scheme, each smart camera is capable of identifying, tracking identified objects, and delivering associate ID information to sibling subjects. So in this paper, we suggest a privacy preservation scheme for preventing privacy infringement during propagation of sensitive information.

Secure discovery method of devices for a home network middleware

With a home network, a device can dynamically join a home network, obtain an IP address, convey its capabilities, and learn about the presence and capabilities of other devices. The devices can subsequently communicate with each other directly. Device discovery protocol defines how network services can be discovered on the network. In this paper, we propose the secure discovery method of devices that uses mutual authentication with symmetric key between devices. This method that we present distributes symmetric key to home network devices by the key distribution server. Using this key, mutual authentication is performed between home appliances. It enables any appliance under any middlewares control to securely communicate any other appliances