Identification of spatially quantised tachistoscopic images of faces: How many pixels does it take to carry identity?

Bachmann, Talis

doi:10.1080/09541449108406221

Cited by 137 publications

(131 citation statements)

References 6 publications

Supporting

Mentioning

110

Contrasting

Order By: Relevance

“…Although images of people certainly do contain a lot of information about superficial fine-scale aspects of the face, if these are not diagnostic of identity, a match without them is likely to improve performance. This notion is, in fact, consistent with research on spatial scale in face recognition, which suggests that identity tends to be carried at low spatial scales (Bachmann, 1991;Harmon & Julesz, 1973; though see Schyns & Oliva, 1997, for an argument that extraction of information from different spatial scales is more flexible when the task is to identify a specific image of a face).…”

Section: Resultssupporting

confidence: 78%

Robust representations for face recognition: The power of averages

Burton

Jenkins

Hancock

et al. 2005

Cognitive Psychology

274

297

View full text Add to dashboard Cite

We are able to recognise familiar faces easily across large variations in image quality, though our ability to match unfamiliar faces is strikingly poor. Here we ask how the representation of a face changes as we become familiar with it. We use a simple imageaveraging technique to derive abstract representations of known faces. Using Principal Components Analysis, we show that computational systems based on these averages consistently outperform systems based on collections of instances. Furthermore, the quality of the average improves as more images are used to derive it. These simulations are carried out with famous faces, over which we had no control of superficial image characteristics. We then present data from three experiments demonstrating that image averaging can also improve recognition by human observers. Finally, we describe how PCA on image averages appears to preserve identity-specific face information, while eliminating non-diagnostic pictorial information. We therefore suggest that this is a good candidate for a robust face representation.2 Robust representations for face recognition: the power of averages

show abstract

Section: Resultssupporting

confidence: 78%

Robust representations for face recognition: The power of averages

Burton

Jenkins

Hancock

et al. 2005

Cognitive Psychology

274

297

View full text Add to dashboard Cite

show abstract

“…In accordance with the preceding results, Bachmann (1991) found similar data when he pixelized faces at different values (15,18,21,24,27,32,44 and 74 pixels/fw) and showed them at 6 different exposure times (1,4,8,20,40 and 100 ms). The percentage of correct identifications increased with the increase in the SFs, but only when the images with 15 pixels/fw were compared with the remaining pixelization conditions, which did not show differences among them.…”

Section: Empirical Evidence Supporting the Hypothesis Of Anisotropic supporting

confidence: 90%

“…During the test phase the subjects were instructed to recognize faces that were shown on the screen for only 100 ms. Each face in the test phase had 11, 21, and 42 pixels (5.5, 10.5 and 21 cycles/fw, respectively), had been low-pass filtered and had been blurred using Gaussian filters. The results indicated a critical value (8 cycles/fw), below which recognition decreased dramatically (similar results were obtained in parametric studies carried out by Bachmann [1991] and Bhatia, Lakshminarayanan, Samal & Welland [1995]). …”

Section: Empirical Evidence Supporting the Critical Role Of Medium-rasupporting

confidence: 83%

Face perception: An integrative review of the role of spatial frequencies

Ruiz-Soler

Beltran

2005

Psychological Research

View full text Add to dashboard Cite

The aim of this article is to reinterpret the results obtained from the research analyzing the role played by spatial frequencies in face perception. Two main working lines have been explored in this body of research: (a) the critical bandwidth of spatial frequencies that allows face recognition to take place (the masking approach) and (b) the role played by different spatial frequencies while the visual percept is being developed (the microgenetic approach). However, results obtained to date are not satisfactory in that no single explanation accounts for all the data obtained from each of the approaches. We propose that the main factor for understanding the role of spatial frequencies in face perception depends on the interaction between the demands of the task and the information in the image (the diagnostic-recognition approach). Using this new framework, we review the most significant research carried out since the early 1970s to provide a reinterpretation of the data obtained. person/machine interfaces in the near future. The importance of this process has motivated the study of these underlying mechanisms for more than three decades. However, after so much effort, there is still no theory that offers a full explanation of all the results obtained.The research aimed at providing an explanation of the face-recognition process is basically focused on three approaches: cognitive, psychophysical and neurophysiological (see Table 1).The cognitive approach has tried to identify the variables that affect successful perceptual tasks (similarity of stimuli, observation time, level of processing, etc.) and to describe as far as possible the different stages in the process in order to generate a high-level symbolic explanatory model

show abstract

“…This can be achieved by following the Rotakin screen height recommendations provided by the Home Offi ce ( Cohen et al ., 2006 ). The video resolution of CCTV video should be adequate so that target faces do not fall below the minimum pixel count of 16 × 16 pixels otherwise identifi cation will be diffi cult (Bachmann, 1991 ). As a general rule of thumb, CCTV cameras in high-risk environments should record and distribute video at very little compression, high resolution and at a frame rate of 12 fps or above.…”

Section: Recommendationsmentioning

confidence: 99%

“Not the Usual Suspects”: A study of factors reducing the effectiveness of CCTV

Keval

Sasse

2008

Secur J

View full text Add to dashboard Cite

Previous research on the effectiveness of Closed Circuit Television (CCTV) has focused on critically assessing police and government claims that CCTV is effective in reducing crime. This paper presents a fi eld study that investigates the relationship between CCTV system design and the performance of operator tasks. We carried out structured observations and interviews with 13 managers and 38 operators at 13 CCTV control rooms. A number of failures were identifi ed, including the poor confi guration of technology, poor quality video recordings, and a lack of system integration. Stakeholder communication was poor, and there were too many cameras and too few operators. These failures have been previously identifi ed by researchers; however, no design improvements have been made to control rooms in the last decade. We identify a number of measures to improve operator performance, and contribute a set of recommendations for security managers and practitioners.

show abstract

Identification of spatially quantised tachistoscopic images of faces: How many pixels does it take to carry identity?

Cited by 137 publications

References 6 publications

Robust representations for face recognition: The power of averages

Robust representations for face recognition: The power of averages

Face perception: An integrative review of the role of spatial frequencies

“Not the Usual Suspects”: A study of factors reducing the effectiveness of CCTV

Contact Info

Product

Resources

About