Hepatic steatosis is one of the most obvious indicators of nonalcoholic fatty liver disease. However, the presence of many regions with a similar color and shape as lipid droplets in the histopathological image complicates the task of detecting genuine lipid droplets using automated methods. Accordingly, the present study proposes a boundary segmentation and matching (BSMatch) algorithm for the segmentation of lipid droplets based on their unique boundary characteristics. A two-branch RnB-Unet model is trained to segment the regions and boundaries of the droplets, respectively, in accordance with a boundary matching (BM) loss which enforces the consistency between them. A boundary matching score (BMS) measure is then used to improve the precision of the instance segmentation evaluation process by discarding segmented regions which are not well-matched with their predicted boundaries. The experimental results obtained using a H&E-stained liver slide dataset show that BSMatch outperforms existing methods in terms of both the IoU and the F1-score. The BSMatch results are used to predict the fat percentage in hepatocytes (FPH) in liver whole slide images. The predicted FPH values are well correlated with the steatosis grades assigned by experienced pathologists. Thus, BSMatch appears to have significant promise for NAFLD diagnosis in clinical contexts.

A channel-and-sensing-aware channel access (CSCA) policy is proposed for multi-channel interweave cognitive radio systems, where multiple secondary users (SUs) are competing for transmission to a common access point. The proposed CSCA policy consists of two key elements: (i) a decentralized channel selection policy that allows each SU to utilize knowledge of both the spectrum occupancy information and its local channel state information to make local channel access decisions and (ii) a channel reservation policy that allows each SU to compete for use of its selected channel by emitting short reservation packets at the beginning of each frame. In the reservation period, a channel-aware splitting algorithm is utilized to resolve collision among SUs that are competing for the same channel. The splitting procedure is optimized with respect to SUs' channel selection policy and ensures that the SU with the best channel quality prevails when collision is resolved. The CSCA policy is derived with the goal of maximizing the SU's throughput subject to a constraint on the probability of collision with primary users (PUs). To satisfy the collision probability constraint, a minimum channel gain threshold is set on each channel to limit the probability that the channel is accessed by SUs. An iterative algorithm is proposed to optimize the parameters in the channel selection and reservation policies, and a low-complexity policy is devised for use in systems with large numbers of channels. The proposed CSCA policies allow SUs to exploit optimally the tradeoff between spectrum availability and channel quality.

We propose a reservation-based channel access policy for multi-channel cognitive radio networks. To enhance the throughput of secondary users (SUs), SUs are allowed to select channels opportunistically according to both the local channel state information (CSI) and the spectrum sensing outcomes. SUs will then compete for the right of transmission on the chosen channel by emitting reservation packets to the access point sequentially according to their local CSI. We further devise a proper threshold on channel gains such that only the SUs whose channel gains are sufficiently high can reserve channels and the interference from SUs to the licensed network can be limited. A channel aware splitting algorithm is adopted to schedule the SU with the highest channel gain to transmit at each time instant. From simulations, the proposed channel access policy outperforms the policies that take into consideration only CSI or sensing outcomes.

Multiuser and cooperative diversity gains have been studied extensively in recent years to improve communication reliability and throughput of wireless networks. Multiuser diversity gains have been exploited by scheduling users with the best channel to transmit in each time slot while cooperative diversity gains have been exploited by having users relay each other's messages to the destination. Most works in the literature consider these advantages separately and in systems that rely on a central scheduler. In this work, we show that the throughput of distributed random access systems can be improved by exploiting both cooperative and multiuser diversity where the transmission probabilities of users are adjusted according to their local channel state information. The fully-loaded stability region is derived as the performance measure of our proposed transmission strategy. The optimal transmission control that maximizes the fully-loaded region is shown to be a threshold-like function. That is, each user shall transmit only when its effective channel is sufficiently reliable. The effective channel in this case is defined as each user's contribution to the weighted sum throughput of the network.

In cooperative systems, users achieve spatial diversity and multihop gains by transmitting packets over multiple independent fading paths provided by their partners. Most previous works on cooperative communications focus on the physical layer aspects such as coding, modulation, and transceiver signal processing techniques. In this work, we study the advantages of user cooperation from a MAC layer perspective and devise queueing strategies to exploit cooperative gains in random access networks. Based on the conventional slotted ALOHA protocol, we propose a simple cooperative transmission mechanism for a two-user cooperative pair. We derive the two-user stability region of the proposed system and show the improvements compared to noncooperative systems. The benefits can be attributed to both physical layer cooperation, where users with good channels may relay for those with bad channels, and MAC layer cooperation, where system parameters can be chosen to enhance cooperation and reduce competition. Then, we extend the proposed strategy to a finite-user system that consists of multiple cooperative pairs. By treating each pair as a single transmission entity, we derive inner bounds for the finite-user stability region and propose a ranking system to characterize the transmission entities' relative tendency of being stable (or unstable).

The impact of imperfect channel state information (CSI) on the transmission control of channel-aware slotted ALOHA networks is studied in this work. By taking into consideration the statistics of the channel estimation error to maximize the achievable stable throughput, we obtain the optimal transmission control policy that determines the transmission probability, rate, and power that should be adopted under each channel state. Specifically, with imperfect CSI, users allocate transmission probabilities to channel states that allow for high probability of successful transmission, which is similar to the case with perfect CSI, but the rate allocation must be performed more conservatively in order to avoid transmission error due to imprecise channel estimates. The policies are first developed for single carrier systems but later extended to multi-carrier systems with maximum per-user power constraints. The asymptotic stable throughput is analyzed for both the single and multiple carrier systems as the number of users goes to infinity. We observe that, with channel estimation error, the proposed transmission control policy that takes into consideration the error statistics may significant increase the throughput compared to strategies that are derived under the perfect CSI assumption.

The impact of imperfect channel state information (CSI) on the transmission control of channel-aware slotted ALOHA networks is studied in this work. By taking into consideration the statistics of the channel estimation error to maximize the achievable stable throughput, we obtain the optimal transmission control policy that determines the transmission probability, rate, and power that should be adopted under different channel states. Specifically, with imperfect CSI, we find that high transmission probabilities should be assigned to channel states that allow for high probability of successful transmission (which may be affected by the estimation errors), but rate allocation must be performed more conservatively (compared to the case with perfect CSI) in order to avoid transmission errors due to imprecise channel estimates. The policies are first developed for single carrier systems and then extended to multicarrier systems with maximum per-user power constraints. The asymptotic maximal stable throughput is analyzed for both the single and multiple carrier systems as the number of users goes to infinity. We observe that, with error in the channel estimate, the proposed transmission control policy that takes into consideration the error statistics may significantly increase the throughput compared to strategies that are derived under the perfect CSI assumption.

In this work, we study the effect of imperfect channel-state information (CSI) on the stability and delay of a two-user channel-aware slotted ALOHA system. We assume that the channel is quantized into two states and channel estimation error may occur between channel states. By taking into account the estimation error, we derive the optimal transmission control function that adjusts the users' transmission probabilities based on the estimated channel state. To evaluate the performance, we derive the stability region and average delay of the system and show the loss in performance due to channel estimation errors.

In this work, we analyze the stability region of the slotted ALOHA random access network with cooperative users. With cooperation, the packets corresponding to each user can be transmitted to the destination through the multiple relaying paths formed by the other users. The diversity introduced by the multi-path relaying reduces significantly the probability of outage. We propose two queuing strategies for the cooperative system. In the first strategy, we assume that each user is equipped with two buffers: a source buffer, to record its own packets, and the relay buffer, to store the packets received from its partner. When both buffers are non-empty, the user chooses randomly among the two buffers based on a predetermined probability. An inner bound of the stability region is derived and is shown to outperform the non-cooperative system. In the second case, we assume that each user utilizes only one buffer to store all packets. The relaying packets are stored at the front of the queue at the cooperative user upon arrival. The second scheme outperforms the non-cooperative case only for certain channel conditions and arrival rates. However, the scheme has low complexity and can be implemented with little changes to the original system.