In the mystery of Liliana Resinovichwho passed away on 14 December 2021 in Trieste, the images from the video surveillance cameras which – investigators hypothesize – would have filmed Lilly on that very day seem increasingly fundamental.
The woman was found dead three weeks later, so those shots are believed to be the last of her alive. But there are those who are not sure whether it is her, first and foremost her brother Sergio Resinovichwhich has repeatedly raised doubts in particular in the filming of the bus of piazzale Gioberti. For their part, the investigators want to get to the bottom of things, so much so that among the 25 points of the investigating judge in the new investigations they aim to “specify on the basis of which criterion the 5-minute gap established in relation to the video surveillance system of the police school in via Damiano Chiesa was determined” and to “check whether the time reported in the videos of the bus camera, called Cam 1, relating to the day of 12/14/2021, corresponds to the actual time or not”.
“I believe that it is very important, as the judge rightly writes, to determine whether the camera in Piazzale Gioberti, in addition to the one in Via Damiano Chiesa, is also on time. If there is almost a 7 minute gap, the investigations could take another direction”, the IT expert tells Il Giornale Sara Capoccitticriminalist and forensic analyst who works for the judicial authority (but not in this particular case) and who founded the Forensically project.
Dr. Capoccitti, how can you recognize a person based on video surveillance footage?
“There are international guidelines on the methodology to be adopted to identify subjects via video surveillance. Frames are extracted from files coming from video surveillance systems and we proceed by biometric traits, in particular those of the face which can make facial recognition useful. This is the preliminary way, but the features cannot always be recovered: due to the poor quality of the images, because the subject is wearing a mask or a neck warmer for example, but above all the distance from the lens. These factors make identification difficult and therefore the analysis proceeds under other aspects, for example the characterizing and individualizing traits”.
What is it about?
“The characterizing traits are, for example, those relating to height, but they do not perfectly identify the identity, because many people can have similar heights. Even if detecting a height well below average, for example 1.50 meters, has an important weight for identification purposes. It is the individualizing traits that are unequivocal, since they are peculiar to a particular person.”
What could they be in the case of Liliana Resinovich?
“The white tuft, but in general it could be a tattoo or another sign of recognition, or patterns relating to gait, such as a limp or a particular somatic movement”.
In fact, in another case, that of Pierina Paganelli, the gait of a man filmed by the surveillance camera is being evaluated.
“Substantially the analysis must start from the individualizing traits and then proceed to the analysis. If these features are not evident from the material in the technician’s possession, we move on to other elements. But the quality of the image is decisive, since the scene is mediated by the device: we do not see reality with our eyes, but through it. And the device may not report all existing data in reality and therefore make it difficult to locate a person. In the case of Pierina, the area under analysis is very small, even if it is conducted by Professor Battiato, one of the most knowledgeable people in this field, and he is attempting an improvement by acquiring the original files. In the Resinovich case there is a great pervasiveness of the IT medium”.
What does it mean?
“This case shows us how the acquisition according to the diktats of digital forensics is fundamental. With the various video surveillance videos, at present, it is not scientifically possible to synchronize the various shots, because we do not know the acquisition method, the metadata of the original file. As far as we know, the files were physically viewed, therefore without a forensic procedure, and a report was drawn up, at least that’s what appears in the file. In the minutes it is written that ‘presumably‘the cameras in via Damiano Chiesa are 5 minutes behind, but the acquisition methodology is not known. Therefore there are no means to synchronize with mathematical certainty the time of the video surveillance of via Damiano Chiesa with that of piazzale Gioberti, because we do not know the origin of the files. If the file has been recovered in the meantime, we will see the result of the analyses. For now the data has no scientific validity, and therefore we don’t even know if that person is the same.”
Is it possible that none of the three people framed (in via San Cilino, in via Damiano Chiesa and in piazzale Gioberti) are the same and above all they are not Liliana Resinovich?
“There are five cones of cameras. Four of them refer to the same video surveillance system, then there is another which is part of a camera mounted on a bus, so it is a moving camera. Forensic identification by images follows very precise rules and the images of Via Damiano Chiesa and those relating to the rubbish bins frame the subject at a very important distance. Furthermore, via Chiesa has a certain slope and the lens has a distortion that is very often present in video surveillance lenses: these details make identification more difficult”.
And then?
“There are even more problems with the subject from Piazzale Gioberti: the pixels are too few to reach a scientifically valid conclusion, furthermore it is not clear the direction that the subject could have taken, because once it leaves the cone of vision, the directions there can be at least two. The data that emerge from the analysis, taken individually, are neutral data, since, if not supported by other criminalistic and therefore contextual data, they do not allow the identification of the walking subject”.
In many TV programs, various experts commented on the shooting. A “Who saw it?” a vfx expert also proposed a correspondence between the person in via Damiano Chiesa and in via Gioberti. But above all, many talked about the possible refraction of light which could have altered the perception of the color of the clothes. Why isn’t that really the case?
“It’s not that it’s not like that. This is why it is important to trace the source of the file you are analyzing. When we acquire footage from a video surveillance system to make it readable by other devices, the format we give it involves compression, i.e. a loss of data, a modification of data that could instead be useful in analysis. For this reason it is important to acquire directly from the video surveillance device and analyze it with forensic software, because the data can also be impoverished in terms of colours”.
In fact, there is the problem of the color of the trousers: in the shot of via Damiano Chiesa they seem light grey, but Liliana Resinovich was found with a pair of very dark trousers, almost black.
“There is a difference in color in the shot between the shoes and the clothes. What the vfx expert did is suggestive and very intuitive, but there is a problem: if I am doing an analysis, I cannot alter the data, I have the duty to keep the data intact throughout the analysis”.
In the second investigations, scientific data seems to prevail: DNA, video cameras, cell phones, contents of smartphones, insects. Will they be able to help find a solution to the case?
“He could have. Because an acquisition according to the rules would have given a better chance of resolution with the acquisition of the various devices, including those for video surveillance or those of her husband Sebastiano Visintin’s GoPro: the GoPro in fact presents a time gap with the tracking. But there are other problems too: too many anomalies at the crime scene to confidently deduce a suicide. Furthermore, the elements appear more attributable to the need to wrap the body, such as the double bag on the head, rather than an intentional alteration of the places, with the aim of staging a suicide.
However, returning to the video surveillance, when it was said that the person in Piazzale Gioberti would have taken the direction of the grove, this is not a certain and proven fact. It is a possibility, there is a difference between formulating a hypothesis and verifying it.”